Authors: Samuel Lavoie, Polina Kirichenko, Mark Ibrahim, Mahmoud Assran, Andrew Gordon Wildon, Aaron Courville, Nicolas Ballas
Abstract: There are a thousand ways to caption an image. Contrastive Language Pretraining (CLIP) on the other hand, works by mapping an image and its caption to a single vector -- limiting how well CLIP-like models can represent the diverse ways to describe an image. In this work, we introduce Llip, Latent Language Image Pretraining, which models the diversity of captions that could match an image. Llip's vision encoder outputs a set of visual features that are mixed into a final representation by conditioning on information derived from the text. We show that Llip outperforms non-contextualized baselines like CLIP and SigLIP on a variety of tasks even with large-scale encoders. Llip improves zero-shot classification by an average of 2.9% zero-shot classification benchmarks with a ViT-G/14 encoder. Specifically, Llip attains a zero-shot top-1 accuracy of 83.5% on ImageNet outperforming a similarly sized CLIP by 1.4%. We also demonstrate improvement on zero-shot retrieval on MS-COCO by 6.0%. We provide a comprehensive analysis of the components introduced by the method and demonstrate that Llip leads to richer visual representations.
Authors: Sicheng Zhao, Hui Chen, Hu Huang, Pengfei Xu, Guiguang Ding
Abstract: In many practical applications, it is often difficult and expensive to obtain large-scale labeled data to train state-of-the-art deep neural networks. Therefore, transferring the learned knowledge from a separate, labeled source domain to an unlabeled or sparsely labeled target domain becomes an appealing alternative. However, direct transfer often results in significant performance decay due to domain shift. Domain adaptation (DA) aims to address this problem by aligning the distributions between the source and target domains. Multi-source domain adaptation (MDA) is a powerful and practical extension in which the labeled data may be collected from multiple sources with different distributions. In this survey, we first define various MDA strategies. Then we systematically summarize and compare modern MDA methods in the deep learning era from different perspectives, followed by commonly used datasets and a brief benchmark. Finally, we discuss future research directions for MDA that are worth investigating.
Authors: Gustavo Adolfo Vargas Hakim, David Osowiechi, Mehrdad Noori, Milad Cheraghalikhani, Ali Bahri, Moslem Yazdanpanah, Ismail Ben Ayed, Christian Desrosiers
Abstract: Pre-trained vision-language models (VLMs), exemplified by CLIP, demonstrate remarkable adaptability across zero-shot classification tasks without additional training. However, their performance diminishes in the presence of domain shifts. In this study, we introduce CLIP Adaptation duRing Test-Time (CLIPArTT), a fully test-time adaptation (TTA) approach for CLIP, which involves automatic text prompts construction during inference for their use as text supervision. Our method employs a unique, minimally invasive text prompt tuning process, wherein multiple predicted classes are aggregated into a single new text prompt, used as pseudo label to re-classify inputs in a transductive manner. Additionally, we pioneer the standardization of TTA benchmarks (e.g., TENT) in the realm of VLMs. Our findings demonstrate that, without requiring additional transformations nor new trainable modules, CLIPArTT enhances performance dynamically across non-corrupted datasets such as CIFAR-10, corrupted datasets like CIFAR-10-C and CIFAR-10.1, alongside synthetic datasets such as VisDA-C. This research underscores the potential for improving VLMs' adaptability through novel test-time strategies, offering insights for robust performance across varied datasets and environments. The code can be found at: https://github.com/dosowiechi/CLIPArTT.git
Authors: Xiaoshi Wu, Yiming Hao, Manyuan Zhang, Keqiang Sun, Zhaoyang Huang, Guanglu Song, Yu Liu, Hongsheng Li
Abstract: Optimizing a text-to-image diffusion model with a given reward function is an important but underexplored research area. In this study, we propose Deep Reward Tuning (DRTune), an algorithm that directly supervises the final output image of a text-to-image diffusion model and back-propagates through the iterative sampling process to the input noise. We find that training earlier steps in the sampling process is crucial for low-level rewards, and deep supervision can be achieved efficiently and effectively by stopping the gradient of the denoising network input. DRTune is extensively evaluated on various reward models. It consistently outperforms other algorithms, particularly for low-level control signals, where all shallow supervision methods fail. Additionally, we fine-tune Stable Diffusion XL 1.0 (SDXL 1.0) model via DRTune to optimize Human Preference Score v2.1, resulting in the Favorable Diffusion XL 1.0 (FDXL 1.0) model. FDXL 1.0 significantly enhances image quality compared to SDXL 1.0 and reaches comparable quality compared with Midjourney v5.2.
Authors: Barak Battash, Amit Rozner, Lior Wolf, Ofir Lindenbaum
Abstract: Large-scale text-to-image models that can generate high-quality and diverse images based on textual prompts have shown remarkable success. These models aim ultimately to create complex scenes, and addressing the challenge of multi-subject generation is a critical step towards this goal. However, the existing state-of-the-art diffusion models face difficulty when generating images that involve multiple subjects. When presented with a prompt containing more than one subject, these models may omit some subjects or merge them together. To address this challenge, we propose a novel approach based on a guiding principle. We allow the diffusion model to initially propose a layout, and then we rearrange the layout grid. This is achieved by enforcing cross-attention maps (XAMs) to adhere to proposed masks and by migrating pixels from latent maps to new locations determined by us. We introduce new loss terms aimed at reducing XAM entropy for clearer spatial definition of subjects, reduce the overlap between XAMs, and ensure that XAMs align with their respective masks. We contrast our approach with several alternative methods and show that it more faithfully captures the desired concepts across a variety of text prompts.
Authors: Shengze Wang, Xueting Li, Chao Liu, Matthew Chan, Michael Stengel, Josef Spjut, Henry Fuchs, Shalini De Mello, Koki Nagano
Abstract: Recent breakthroughs in single-image 3D portrait reconstruction have enabled telepresence systems to stream 3D portrait videos from a single camera in real-time, potentially democratizing telepresence. However, per-frame 3D reconstruction exhibits temporal inconsistency and forgets the user's appearance. On the other hand, self-reenactment methods can render coherent 3D portraits by driving a personalized 3D prior, but fail to faithfully reconstruct the user's per-frame appearance (e.g., facial expressions and lighting). In this work, we recognize the need to maintain both coherent identity and dynamic per-frame appearance to enable the best possible realism. To this end, we propose a new fusion-based method that fuses a personalized 3D subject prior with per-frame information, producing temporally stable 3D videos with faithful reconstruction of the user's per-frame appearances. Trained only using synthetic data produced by an expression-conditioned 3D GAN, our encoder-based method achieves both state-of-the-art 3D reconstruction accuracy and temporal consistency on in-studio and in-the-wild datasets.
Authors: Hui Lin, Charilaos Apostolidis, Aggelos K. Katsaggelos
Abstract: Differences in image quality, lighting conditions, and patient demographics pose challenges to automated glaucoma detection from color fundus photography. Brighteye, a method based on Vision Transformer, is proposed for glaucoma detection and glaucomatous feature classification. Brighteye learns long-range relationships among pixels within large fundus images using a self-attention mechanism. Prior to being input into Brighteye, the optic disc is localized using YOLOv8, and the region of interest (ROI) around the disc center is cropped to ensure alignment with clinical practice. Optic disc detection improves the sensitivity at 95% specificity from 79.20% to 85.70% for glaucoma detection and the Hamming distance from 0.2470 to 0.1250 for glaucomatous feature classification. In the developmental stage of the Justified Referral in AI Glaucoma Screening (JustRAIGS) challenge, the overall outcome secured the fifth position out of 226 entries.
Authors: Palawat Busaranuvong, Emmanuel Agu, Deepak Kumar, Shefalika Gautam, Reza Saadati Fard, Bengisu Tulu, Diane Strong
Abstract: To detect infected wounds in Diabetic Foot Ulcers (DFUs) from photographs, preventing severe complications and amputations. Methods: This paper proposes the Guided Conditional Diffusion Classifier (ConDiff), a novel deep-learning infection detection model that combines guided image synthesis with a denoising diffusion model and distance-based classification. The process involves (1) generating guided conditional synthetic images by injecting Gaussian noise to a guide image, followed by denoising the noise-perturbed image through a reverse diffusion process, conditioned on infection status and (2) classifying infections based on the minimum Euclidean distance between synthesized images and the original guide image in embedding space. Results: ConDiff demonstrated superior performance with an accuracy of 83% and an F1-score of 0.858, outperforming state-of-the-art models by at least 3%. The use of a triplet loss function reduces overfitting in the distance-based classifier. Conclusions: ConDiff not only enhances diagnostic accuracy for DFU infections but also pioneers the use of generative discriminative models for detailed medical image analysis, offering a promising approach for improving patient outcomes.
Authors: Prateek Verma, Minh-Hao Van, Xintao Wu
Abstract: Vision language models (VLMs) have recently emerged and gained the spotlight for their ability to comprehend the dual modality of image and textual data. VLMs such as LLaVA, ChatGPT-4, and Gemini have recently shown impressive performance on tasks such as natural image captioning, visual question answering (VQA), and spatial reasoning. Additionally, a universal segmentation model by Meta AI, Segment Anything Model (SAM) shows unprecedented performance at isolating objects from unforeseen images. Since medical experts, biologists, and materials scientists routinely examine microscopy or medical images in conjunction with textual information in the form of captions, literature, or reports, and draw conclusions of great importance and merit, it is indubitably essential to test the performance of VLMs and foundation models such as SAM, on these images. In this study, we charge ChatGPT, LLaVA, Gemini, and SAM with classification, segmentation, counting, and VQA tasks on a variety of microscopy images. We observe that ChatGPT and Gemini are impressively able to comprehend the visual features in microscopy images, while SAM is quite capable at isolating artefacts in a general sense. However, the performance is not close to that of a domain expert - the models are readily encumbered by the introduction of impurities, defects, artefact overlaps and diversity present in the images.
Authors: Burak Can Biner, Farrin Marouf Sofian, Umur Berkay Karaka\c{s}, Duygu Ceylan, Erkut Erdem, Aykut Erdem
Abstract: We are witnessing a revolution in conditional image synthesis with the recent success of large scale text-to-image generation methods. This success also opens up new opportunities in controlling the generation and editing process using multi-modal input. While spatial control using cues such as depth, sketch, and other images has attracted a lot of research, we argue that another equally effective modality is audio since sound and sight are two main components of human perception. Hence, we propose a method to enable audio-conditioning in large scale image diffusion models. Our method first maps features obtained from audio clips to tokens that can be injected into the diffusion model in a fashion similar to text tokens. We introduce additional audio-image cross attention layers which we finetune while freezing the weights of the original layers of the diffusion model. In addition to audio conditioned image generation, our method can also be utilized in conjuction with diffusion based editing methods to enable audio conditioned image editing. We demonstrate our method on a wide range of audio and image datasets. We perform extensive comparisons with recent methods and show favorable performance.
Authors: Colby Banbury, Emil Njor, Matthew Stewart, Pete Warden, Manjunath Kudlur, Nat Jeffries, Xenofon Fafoutis, Vijay Janapa Reddi
Abstract: Machine learning applications on extremely low-power devices, commonly referred to as tiny machine learning (TinyML), promises a smarter and more connected world. However, the advancement of current TinyML research is hindered by the limited size and quality of pertinent datasets. To address this challenge, we introduce Wake Vision, a large-scale, diverse dataset tailored for person detection -- the canonical task for TinyML visual sensing. Wake Vision comprises over 6 million images, which is a hundredfold increase compared to the previous standard, and has undergone thorough quality filtering. Using Wake Vision for training results in a 2.41\% increase in accuracy compared to the established benchmark. Alongside the dataset, we provide a collection of five detailed benchmark sets that assess model performance on specific segments of the test data, such as varying lighting conditions, distances from the camera, and demographic characteristics of subjects. These novel fine-grained benchmarks facilitate the evaluation of model quality in challenging real-world scenarios that are often ignored when focusing solely on overall accuracy. Through an evaluation of a MobileNetV2 TinyML model on the benchmarks, we show that the input resolution plays a more crucial role than the model width in detecting distant subjects and that the impact of quantization on model robustness is minimal, thanks to the dataset quality. These findings underscore the importance of a detailed evaluation to identify essential factors for model development. The dataset, benchmark suite, code, and models are publicly available under the CC-BY 4.0 license, enabling their use for commercial use cases.
Authors: Shanlin Sun, Bingbing Zhuang, Ziyu Jiang, Buyu Liu, Xiaohui Xie, Manmohan Chandraker
Abstract: Photorealistic simulation plays a crucial role in applications such as autonomous driving, where advances in neural radiance fields (NeRFs) may allow better scalability through the automatic creation of digital 3D assets. However, reconstruction quality suffers on street scenes due to largely collinear camera motions and sparser samplings at higher speeds. On the other hand, the application often demands rendering from camera views that deviate from the inputs to accurately simulate behaviors like lane changes. In this paper, we propose several insights that allow a better utilization of Lidar data to improve NeRF quality on street scenes. First, our framework learns a geometric scene representation from Lidar, which is fused with the implicit grid-based representation for radiance decoding, thereby supplying stronger geometric information offered by explicit point cloud. Second, we put forth a robust occlusion-aware depth supervision scheme, which allows utilizing densified Lidar points by accumulation. Third, we generate augmented training views from Lidar points for further improvement. Our insights translate to largely improved novel view synthesis under real driving scenes.
Authors: Ojasw Upadhyay
Abstract: Vision transformers have revolutionized computer vision, but their computational demands present challenges for training and deployment. This paper introduces LOTUS (LOttery Transformers with Ultra Sparsity), a novel method that leverages data lottery ticket selection and sparsity pruning to accelerate vision transformer training while maintaining accuracy. Our approach focuses on identifying and utilizing the most informative data subsets and eliminating redundant model parameters to optimize the training process. Through extensive experiments, we demonstrate the effectiveness of LOTUS in achieving rapid convergence and high accuracy with significantly reduced computational requirements. This work highlights the potential of combining data selection and sparsity techniques for efficient vision transformer training, opening doors for further research and development in this area.
Authors: K. Yeh, M. S. Jabal, V. Gupta, D. F. Kallmes, W. Brinjikji, B. S. Erdal
Abstract: Background and Purpose: Identifying the thromboembolism source in ischemic stroke is crucial for treatment and secondary prevention yet is often undetermined. This study describes a self-supervised deep learning approach in digital pathology of emboli for classifying ischemic stroke clot origin from histopathological images. Methods: The dataset included whole slide images (WSI) from the STRIP AI Kaggle challenge, consisting of retrieved clots from ischemic stroke patients following mechanical thrombectomy. Transformer-based deep learning models were developed using transfer learning and self-supervised pretraining for classifying WSI. Customizations included an attention pooling layer, weighted loss function, and threshold optimization. Various model architectures were tested and compared, and model performances were primarily evaluated using weighted logarithmic loss. Results: The model achieved a logloss score of 0.662 in cross-validation and 0.659 on the test set. Different model backbones were compared, with the swin_large_patch4_window12_384 showed higher performance. Thresholding techniques for clot origin classification were employed to balance false positives and negatives. Conclusion: The study demonstrates the extent of efficacy of transformer-based deep learning models in identifying ischemic stroke clot origins from histopathological images and emphasizes the need for refined modeling techniques specifically adapted to thrombi WSI. Further research is needed to improve model performance, interpretability, validate its effectiveness. Future enhancement could include integrating larger patient cohorts, advanced preprocessing strategies, and exploring ensemble multimodal methods for enhanced diagnostic accuracy.
Authors: Guangyao Zhai, Evin P{\i}nar \"Ornek, Dave Zhenyu Chen, Ruotong Liao, Yan Di, Nassir Navab, Federico Tombari, Benjamin Busam
Abstract: We present EchoScene, an interactive and controllable generative model that generates 3D indoor scenes on scene graphs. EchoScene leverages a dual-branch diffusion model that dynamically adapts to scene graphs. Existing methods struggle to handle scene graphs due to varying numbers of nodes, multiple edge combinations, and manipulator-induced node-edge operations. EchoScene overcomes this by associating each node with a denoising process and enables collaborative information exchange, enhancing controllable and consistent generation aware of global constraints. This is achieved through an information echo scheme in both shape and layout branches. At every denoising step, all processes share their denoising data with an information exchange unit that combines these updates using graph convolution. The scheme ensures that the denoising processes are influenced by a holistic understanding of the scene graph, facilitating the generation of globally coherent scenes. The resulting scenes can be manipulated during inference by editing the input scene graph and sampling the noise in the diffusion model. Extensive experiments validate our approach, which maintains scene controllability and surpasses previous methods in generation fidelity. Moreover, the generated scenes are of high quality and thus directly compatible with off-the-shelf texture generation. Code and trained models are open-sourced.
Authors: Somesh Singh, Harini S I, Yaman K Singla, Veeky Baths, Rajiv Ratn Shah, Changyou Chen, Balaji Krishnamurthy
Abstract: Communication is defined as ``Who says what to whom with what effect.'' A message from a communicator generates downstream receiver effects, also known as behavior. Receiver behavior, being a downstream effect of the message, carries rich signals about it. Even after carrying signals about the message, the behavior data is often ignored while training large language models. We show that training LLMs on receiver behavior can actually help improve their content-understanding abilities. Specifically, we show that training LLMs to predict the receiver behavior of likes and comments improves the LLM's performance on a wide variety of downstream content understanding tasks. We show this performance increase over 40 video and image understanding tasks over 23 benchmark datasets across both 0-shot and fine-tuning settings, outperforming many supervised baselines. Moreover, since receiver behavior, such as likes and comments, is collected by default on the internet and does not need any human annotations to be useful, the performance improvement we get after training on this data is essentially free-lunch. We release the receiver behavior cleaned comments and likes of 750k images and videos collected from multiple platforms along with our instruction-tuning data.
Authors: Katherine Henneberger, Jing Qin
Abstract: Hyperspectral Imaging (HSI) serves as an important technique in remote sensing. However, high dimensionality and data volume typically pose significant computational challenges. Band selection is essential for reducing spectral redundancy in hyperspectral imagery while retaining intrinsic critical information. In this work, we propose a novel hyperspectral band selection model by decomposing the data into a low-rank and smooth component and a sparse one. In particular, we develop a generalized 3D total variation (G3DTV) by applying the $\ell_1^p$-norm to derivatives to preserve spatial-spectral smoothness. By employing the alternating direction method of multipliers (ADMM), we derive an efficient algorithm, where the tensor low-rankness is implied by the tensor CUR decomposition. We demonstrate the effectiveness of the proposed approach through comparisons with various other state-of-the-art band selection techniques using two benchmark real-world datasets. In addition, we provide practical guidelines for parameter selection in both noise-free and noisy scenarios.
Authors: Yiwei Ma, Zhekai Lin, Jiayi Ji, Yijun Fan, Xiaoshuai Sun, Rongrong Ji
Abstract: Recent advancements in automatic 3D avatar generation guided by text have made significant progress. However, existing methods have limitations such as oversaturation and low-quality output. To address these challenges, we propose X-Oscar, a progressive framework for generating high-quality animatable avatars from text prompts. It follows a sequential Geometry->Texture->Animation paradigm, simplifying optimization through step-by-step generation. To tackle oversaturation, we introduce Adaptive Variational Parameter (AVP), representing avatars as an adaptive distribution during training. Additionally, we present Avatar-aware Score Distillation Sampling (ASDS), a novel technique that incorporates avatar-aware noise into rendered images for improved generation quality during optimization. Extensive evaluations confirm the superiority of X-Oscar over existing text-to-3D and text-to-avatar approaches. Our anonymous project page: https://xmu-xiaoma666.github.io/Projects/X-Oscar/.
Authors: Honglong Yang, Hui Tang, Xiaomeng Li
Abstract: Radiology report generation aims to automatically generate detailed and coherent descriptive reports alongside radiology images. Previous work mainly focused on refining fine-grained image features or leveraging external knowledge. However, the precise alignment of fine-grained image features with corresponding text descriptions has not been considered. This paper presents a novel method called Fine-grained Image-Text Aligner (FITA) to construct fine-grained alignment for image and text features. It has three novel designs: Image Feature Refiner (IFR), Text Feature Refiner (TFR) and Contrastive Aligner (CA). IFR and TFR aim to learn fine-grained image and text features, respectively. We achieve this by leveraging saliency maps to effectively fuse symptoms with corresponding abnormal visual regions, and by utilizing a meticulously constructed triplet set for training. Finally, CA module aligns fine-grained image and text features using contrastive loss for precise alignment. Results show that our method surpasses existing methods on the widely used benchmark
Authors: Peng Chu, Jiang Wang, Andre Abrantes
Abstract: The development of Audio Description (AD) has been a pivotal step forward in making video content more accessible and inclusive. Traditionally, AD production has demanded a considerable amount of skilled labor, while existing automated approaches still necessitate extensive training to integrate multimodal inputs and tailor the output from a captioning style to an AD style. In this paper, we introduce an automated AD generation pipeline that harnesses the potent multimodal and instruction-following capacities of GPT-4V(ision). Notably, our methodology employs readily available components, eliminating the need for additional training. It produces ADs that not only comply with established natural language AD production standards but also maintain contextually consistent character information across frames, courtesy of a tracking-based character recognition module. A thorough analysis on the MAD dataset reveals that our approach achieves a performance on par with learning-based methods in automated AD production, as substantiated by a CIDEr score of 20.5.
Authors: Xin Yan
Abstract: Building height is an important indicator for scientific research and practical application. However, building height products with a high spatial resolution (10m) are still very scarce. To meet the needs of high-resolution building height estimation models, this study established a set of spatial-spectral-temporal feature databases, combining SAR data provided by Sentinel-1, optical data provided by Sentinel-2, and shape data provided by building footprints. The statistical indicators on the time scale are extracted to form a rich database of 160 features. This study combined with permutation feature importance, Shapley Additive Explanations, and Random Forest variable importance, and the final stable features are obtained through an expert scoring system. This study took 12 large, medium, and small cities in the United States as the training data. It used moving windows to aggregate the pixels to solve the impact of SAR image displacement and building shadows. This study built a building height model based on a random forest model and compared three model ensemble methods of bagging, boosting, and stacking. To evaluate the accuracy of the prediction results, this study collected Lidar data in the test area, and the evaluation results showed that its R-Square reached 0.78, which can prove that the building height can be obtained effectively. The fast production of high-resolution building height data can support large-scale scientific research and application in many fields.
Authors: Yuhang Huang, SHilong Zou, Xinwang Liu, Kai Xu
Abstract: This paper presents a novel latent 3D diffusion model for the generation of neural voxel fields, aiming to achieve accurate part-aware structures. Compared to existing methods, there are two key designs to ensure high-quality and accurate part-aware generation. On one hand, we introduce a latent 3D diffusion process for neural voxel fields, enabling generation at significantly higher resolutions that can accurately capture rich textural and geometric details. On the other hand, a part-aware shape decoder is introduced to integrate the part codes into the neural voxel fields, guiding the accurate part decomposition and producing high-quality rendering results. Through extensive experimentation and comparisons with state-of-the-art methods, we evaluate our approach across four different classes of data. The results demonstrate the superior generative capabilities of our proposed method in part-aware shape generation, outperforming existing state-of-the-art methods.
Authors: Xiaoqi Zhao, Youwei Pang, Wei Ji, Baicheng Sheng, Jiaming Zuo, Lihe Zhang, Huchuan Lu
Abstract: Different from the context-independent (CI) concepts such as human, car, and airplane, context-dependent (CD) concepts require higher visual understanding ability, such as camouflaged object and medical lesion. Despite the rapid advance of many CD understanding tasks in respective branches, the isolated evolution leads to their limited cross-domain generalisation and repetitive technique innovation. Since there is a strong coupling relationship between foreground and background context in CD tasks, existing methods require to train separate models in their focused domains. This restricts their real-world CD concept understanding towards artificial general intelligence (AGI). We propose a unified model with a single set of parameters, Spider, which only needs to be trained once. With the help of the proposed concept filter driven by the image-mask group prompt, Spider is able to understand and distinguish diverse strong context-dependent concepts to accurately capture the Prompter's intention. Without bells and whistles, Spider significantly outperforms the state-of-the-art specialized models in 8 different context-dependent segmentation tasks, including 4 natural scenes (salient, camouflaged, and transparent objects and shadow) and 4 medical lesions (COVID-19, polyp, breast, and skin lesion with color colonoscopy, CT, ultrasound, and dermoscopy modalities). Besides, Spider shows obvious advantages in continuous learning. It can easily complete the training of new tasks by fine-tuning parameters less than 1\% and bring a tolerable performance degradation of less than 5\% for all old tasks. The source code will be publicly available at \href{https://github.com/Xiaoqi-Zhao-DLUT/Spider-UniCDSeg}{Spider-UniCDSeg}.
Authors: Samyadeep Basu, Keivan Rezaei, Ryan Rossi, Cherry Zhao, Vlad Morariu, Varun Manjunatha, Soheil Feizi
Abstract: Identifying layers within text-to-image models which control visual attributes can facilitate efficient model editing through closed-form updates. Recent work, leveraging causal tracing show that early Stable-Diffusion variants confine knowledge primarily to the first layer of the CLIP text-encoder, while it diffuses throughout the UNet.Extending this framework, we observe that for recent models (e.g., SD-XL, DeepFloyd), causal tracing fails in pinpointing localized knowledge, highlighting challenges in model editing. To address this issue, we introduce the concept of Mechanistic Localization in text-to-image models, where knowledge about various visual attributes (e.g., ``style", ``objects", ``facts") can be mechanistically localized to a small fraction of layers in the UNet, thus facilitating efficient model editing. We localize knowledge using our method LocoGen which measures the direct effect of intermediate layers to output generation by performing interventions in the cross-attention layers of the UNet. We then employ LocoEdit, a fast closed-form editing method across popular open-source text-to-image models (including the latest SD-XL)and explore the possibilities of neuron-level model editing. Using Mechanistic Localization, our work offers a better view of successes and failures in localization-based text-to-image model editing. Code will be available at \href{https://github.com/samyadeepbasu/LocoGen}{https://github.com/samyadeepbasu/LocoGen}.
URLs: https://github.com/samyadeepbasu/LocoGen, https://github.com/samyadeepbasu/LocoGen
Authors: Minsu Kim, Giseop Kim, Sunwook Choi
Abstract: Recent advancements in Bird's Eye View (BEV) fusion for map construction have demonstrated remarkable mapping of urban environments. However, their deep and bulky architecture incurs substantial amounts of backpropagation memory and computing latency. Consequently, the problem poses an unavoidable bottleneck in constructing high-resolution (HR) BEV maps, as their large-sized features cause significant increases in costs including GPU memory consumption and computing latency, named diverging training costs issue. Affected by the problem, most existing methods adopt low-resolution (LR) BEV and struggle to estimate the precise locations of urban scene components like road lanes, and sidewalks. As the imprecision leads to risky self-driving, the diverging training costs issue has to be resolved. In this paper, we address the issue with our novel Trumpet Neural Network (TNN) mechanism. The framework utilizes LR BEV space and outputs an up-sampled semantic BEV map to create a memory-efficient pipeline. To this end, we introduce Local Restoration of BEV representation. Specifically, the up-sampled BEV representation has severely aliased, blocky signals, and thick semantic labels. Our proposed Local Restoration restores the signals and thins (or narrows down) the width of the labels. Our extensive experiments show that the TNN mechanism provides a plug-and-play memory-efficient pipeline, thereby enabling the effective estimation of real-sized (or precise) semantic labels for BEV map construction.
Authors: Kiyoon Jeong, Woojun Lee, Woongchan Nam, Minjeong Ma, Pilsung Kang
Abstract: This report presents the ECO (Ensembled Clip score and cOnsensus score) pipeline from team DSBA LAB, which is a new framework used to evaluate and rank captions for a given image. ECO selects the most accurate caption describing image. It is made possible by combining an Ensembled CLIP score, which considers the semantic alignment between the image and captions, with a Consensus score that accounts for the essentialness of the captions. Using this framework, we achieved notable success in the CVPR 2024 Workshop Challenge on Caption Re-ranking Evaluation at the New Frontiers for Zero-Shot Image Captioning Evaluation (NICE). Specifically, we secured third place based on the CIDEr metric, second in both the SPICE and METEOR metrics, and first in the ROUGE-L and all BLEU Score metrics. The code and configuration for the ECO framework are available at https://github.com/ DSBA-Lab/ECO .
URLs: https://github.com/
Authors: Anurag Kumar, Chinmay Bharti, Saikat Dutta, Srikrishna Karanam, Biplab Banerjee
Abstract: Recent advancements in deep learning have demonstrated remarkable performance comparable to human capabilities across various supervised computer vision tasks. However, the prevalent assumption of having an extensive pool of training data encompassing all classes prior to model training often diverges from real-world scenarios, where limited data availability for novel classes is the norm. The challenge emerges in seamlessly integrating new classes with few samples into the training data, demanding the model to adeptly accommodate these additions without compromising its performance on base classes. To address this exigency, the research community has introduced several solutions under the realm of few-shot class incremental learning (FSCIL). In this study, we introduce an innovative FSCIL framework that utilizes language regularizer and subspace regularizer. During base training, the language regularizer helps incorporate semantic information extracted from a Vision-Language model. The subspace regularizer helps in facilitating the model's acquisition of nuanced connections between image and text semantics inherent to base classes during incremental training. Our proposed framework not only empowers the model to embrace novel classes with limited data, but also ensures the preservation of performance on base classes. To substantiate the efficacy of our approach, we conduct comprehensive experiments on three distinct FSCIL benchmarks, where our framework attains state-of-the-art performance.
Authors: Zhenyang Huang, Zhaojin Fu, Song Jintao, Genji Yuan, Jinjiang Li
Abstract: Change detection as an interdisciplinary discipline in the field of computer vision and remote sensing at present has been receiving extensive attention and research. Due to the rapid development of society, the geographic information captured by remote sensing satellites is changing faster and more complex, which undoubtedly poses a higher challenge and highlights the value of change detection tasks. We propose MFDS-Net: Multi-Scale Feature Depth-Supervised Network for Remote Sensing Change Detection with Global Semantic and Detail Information (MFDS-Net) with the aim of achieving a more refined description of changing buildings as well as geographic information, enhancing the localisation of changing targets and the acquisition of weak features. To achieve the research objectives, we use a modified ResNet_34 as backbone network to perform feature extraction and DO-Conv as an alternative to traditional convolution to better focus on the association between feature information and to obtain better training results. We propose the Global Semantic Enhancement Module (GSEM) to enhance the processing of high-level semantic information from a global perspective. The Differential Feature Integration Module (DFIM) is proposed to strengthen the fusion of different depth feature information, achieving learning and extraction of differential features. The entire network is trained and optimized using a deep supervision mechanism. The experimental outcomes of MFDS-Net surpass those of current mainstream change detection networks. On the LEVIR dataset, it achieved an F1 score of 91.589 and IoU of 84.483, on the WHU dataset, the scores were F1: 92.384 and IoU: 86.807, and on the GZ-CD dataset, the scores were F1: 86.377 and IoU: 76.021. The code is available at https://github.com/AOZAKIiii/MFDS-Net
Authors: Zixun Jiao, Xihan Wang, Quanli Gao
Abstract: Reconstructing a hand mesh from a single RGB image is a challenging task because hands are often occluded by objects. Most previous works attempted to introduce more additional information and adopt attention mechanisms to improve 3D reconstruction results, but it would increased computational complexity. This observation prompts us to propose a new and concise architecture while improving computational efficiency. In this work, we propose a simple and effective 3D hand mesh reconstruction network HandSSCA, which is the first to incorporate state space modeling into the field of hand pose estimation. In the network, we have designed a novel state space channel attention module that extends the effective sensory field, extracts hand features in the spatial dimension, and enhances hand regional features in the channel dimension. This design helps to reconstruct a complete and detailed hand mesh. Extensive experiments conducted on well-known datasets featuring challenging hand-object occlusions (such as FREIHAND, DEXYCB, and HO3D) demonstrate that our proposed HandSSCA achieves state-of-the-art performance while maintaining a minimal parameter count.
Authors: Christopher Kermorvant, Eva Bardou, Manon Blanco, Bastien Abadie
Abstract: This paper presents Callico, a web-based open source platform designed to simplify the annotation process in document recognition projects. The move towards data-centric AI in machine learning and deep learning underscores the importance of high-quality data, and the need for specialised tools that increase the efficiency and effectiveness of generating such data. For document image annotation, Callico offers dual-display annotation for digitised documents, enabling simultaneous visualisation and annotation of scanned images and text. This capability is critical for OCR and HTR model training, document layout analysis, named entity recognition, form-based key value annotation or hierarchical structure annotation with element grouping. The platform supports collaborative annotation with versatile features backed by a commitment to open source development, high-quality code standards and easy deployment via Docker. Illustrative use cases - including the transcription of the Belfort municipal registers, the indexing of French World War II prisoners for the ICRC, and the extraction of personal information from the Socface project's census lists - demonstrate Callico's applicability and utility.
Authors: Nianzu Qiao, Lamei Di, Changyin Sun
Abstract: Deep learning-based motion deblurring techniques have advanced significantly in recent years. This class of techniques, however, does not carefully examine the inherent flaws in blurry images. For instance, low edge and structural information are traits of blurry images. The high-frequency component of blurry images is edge information, and the low-frequency component is structure information. A blind motion deblurring network (MCMS) based on multi-category information and multi-scale stripe attention mechanism is proposed. Given the respective characteristics of the high-frequency and low-frequency components, a three-stage encoder-decoder model is designed. Specifically, the first stage focuses on extracting the features of the high-frequency component, the second stage concentrates on extracting the features of the low-frequency component, and the third stage integrates the extracted low-frequency component features, the extracted high-frequency component features, and the original blurred image in order to recover the final clear image. As a result, the model effectively improves motion deblurring by fusing the edge information of the high-frequency component and the structural information of the low-frequency component. In addition, a grouped feature fusion technique is developed so as to achieve richer, more three-dimensional and comprehensive utilization of various types of features at a deep level. Next, a multi-scale stripe attention mechanism (MSSA) is designed, which effectively combines the anisotropy and multi-scale information of the image, a move that significantly enhances the capability of the deep model in feature representation. Large-scale comparative studies on various datasets show that the strategy in this paper works better than the recently published measures.
Authors: Nianzu Qiao, Lamei Di, Changyin Sun
Abstract: Although several image super-resolution solutions exist, they still face many challenges. CNN-based algorithms, despite the reduction in computational complexity, still need to improve their accuracy. While Transformer-based algorithms have higher accuracy, their ultra-high computational complexity makes them difficult to be accepted in practical applications. To overcome the existing challenges, a novel super-resolution reconstruction algorithm is proposed in this paper. The algorithm achieves a significant increase in accuracy through a unique design while maintaining a low complexity. The core of the algorithm lies in its cleverly designed Global-Local Information Extraction Module and Basic Block Module. By combining global and local information, the Global-Local Information Extraction Module aims to understand the image content more comprehensively so as to recover the global structure and local details in the image more accurately, which provides rich information support for the subsequent reconstruction process. Experimental results show that the comprehensive performance of the algorithm proposed in this paper is optimal, providing an efficient and practical new solution in the field of super-resolution reconstruction.
Authors: Nahuel Gonz\'alez, Giuseppe Stragapede, Rub\'en Vera-Rodriguez, Rub\'en Tolosana
Abstract: In 2021, the pioneering work on TypeNet showed that keystroke dynamics verification could scale to hundreds of thousands of users with minimal performance degradation. Recently, the KVC-onGoing competition has provided an open and robust experimental protocol for evaluating keystroke dynamics verification systems of such scale, including considerations of algorithmic fairness. This article describes Type2Branch, the model and techniques that achieved the lowest error rates at the KVC-onGoing, in both desktop and mobile scenarios. The novelty aspects of the proposed Type2Branch include: i) synthesized timing features emphasizing user behavior deviation from the general population, ii) a dual-branch architecture combining recurrent and convolutional paths with various attention mechanisms, iii) a new loss function named Set2set that captures the global structure of the embedding space, and iv) a training curriculum of increasing difficulty. Considering five enrollment samples per subject of approximately 50 characters typed, the proposed Type2Branch achieves state-of-the-art performance with mean per-subject EERs of 0.77% and 1.03% on evaluation sets of respectively 15,000 and 5,000 subjects for desktop and mobile scenarios. With a uniform global threshold for all subjects, the EERs are 3.25% for desktop and 3.61% for mobile, outperforming previous approaches by a significant margin.
Authors: Masatoshi Tateno, Takuma Yagi, Ryosuke Furuta, Yoichi Sato
Abstract: Temporally localizing the presence of object states in videos is crucial in understanding human activities beyond actions and objects. This task has suffered from a lack of training data due to object states' inherent ambiguity and variety. To avoid exhaustive annotation, learning from transcribed narrations in instructional videos would be intriguing. However, object states are less described in narrations compared to actions, making them less effective. In this work, we propose to extract the object state information from action information included in narrations, using large language models (LLMs). Our observation is that LLMs include world knowledge on the relationship between actions and their resulting object states, and can infer the presence of object states from past action sequences. The proposed LLM-based framework offers flexibility to generate plausible pseudo-object state labels against arbitrary categories. We evaluate our method with our newly collected Multiple Object States Transition (MOST) dataset including dense temporal annotation of 60 object state categories. Our model trained by the generated pseudo-labels demonstrates significant improvement of over 29% in mAP against strong zero-shot vision-language models, showing the effectiveness of explicitly extracting object state information from actions through LLMs.
Authors: Muhammad Ahmad, Manuel Mazzara, Salvatore Distifano
Abstract: 3D Swin Transformer (3D-ST) known for its hierarchical attention and window-based processing, excels in capturing intricate spatial relationships within images. Spatial-spectral Transformer (SST), meanwhile, specializes in modeling long-range dependencies through self-attention mechanisms. Therefore, this paper introduces a novel method: an attentional fusion of these two transformers to significantly enhance the classification performance of Hyperspectral Images (HSIs). What sets this approach apart is its emphasis on the integration of attentional mechanisms from both architectures. This integration not only refines the modeling of spatial and spectral information but also contributes to achieving more precise and accurate classification results. The experimentation and evaluation of benchmark HSI datasets underscore the importance of employing disjoint training, validation, and test samples. The results demonstrate the effectiveness of the fusion approach, showcasing its superiority over traditional methods and individual transformers. Incorporating disjoint samples enhances the robustness and reliability of the proposed methodology, emphasizing its potential for advancing hyperspectral image classification.
Authors: Quang-Huy Che, Le-Chuong Nguyen, Vinh-Tiep Nguyen
Abstract: The quest for robust Person re-identification (Re-ID) systems capable of accurately identifying subjects across diverse scenarios remains a formidable challenge in surveillance and security applications. This study presents a novel methodology that significantly enhances Person Re-Identification (Re-ID) by integrating Uncertainty Feature Fusion (UFFM) with Wise Distance Aggregation (WDA). Tested on benchmark datasets - Market-1501, DukeMTMC-ReID, and MSMT17 - our approach demonstrates substantial improvements in Rank-1 accuracy and mean Average Precision (mAP). Specifically, UFFM capitalizes on the power of feature synthesis from multiple images to overcome the limitations imposed by the variability of subject appearances across different views. WDA further refines the process by intelligently aggregating similarity metrics, thereby enhancing the system's ability to discern subtle but critical differences between subjects. The empirical results affirm the superiority of our method over existing approaches, achieving new performance benchmarks across all evaluated datasets. Code is available on Github.
Authors: Matthias Streller, So\v{n}a Michl\'ikov\'a, Willy Ciecior, Katharina L\"onnecke, Leoni A. Kunz-Schughart, Steffen Lange, Anja Voss-B\"ohme
Abstract: Multicellular tumor spheroids (MCTS) are advanced cell culture systems for assessing the impact of combinatorial radio(chemo)therapy. They exhibit therapeutically relevant in-vivo-like characteristics from 3D cell-cell and cell-matrix interactions to radial pathophysiological gradients related to proliferative activity and nutrient/oxygen supply, altering cellular radioresponse. State-of-the-art assays quantify long-term curative endpoints based on collected brightfield image time series from large treated spheroid populations per irradiation dose and treatment arm. Here, spheroid control probabilities are documented analogous to in-vivo tumor control probabilities based on Kaplan-Meier curves. This analyses require laborious spheroid segmentation of up to 100.000 images per treatment arm to extract relevant structural information from the images, e.g., diameter, area, volume and circularity. While several image analysis algorithms are available for spheroid segmentation, they all focus on compact MCTS with clearly distinguishable outer rim throughout growth. However, treated MCTS may partly be detached and destroyed and are usually obscured by dead cell debris. We successfully train two Fully Convolutional Networks, UNet and HRNet, and optimize their hyperparameters to develop an automatic segmentation for both untreated and treated MCTS. We systematically validate the automatic segmentation on larger, independent data sets of spheroids derived from two human head-and-neck cancer cell lines. We find an excellent overlap between manual and automatic segmentation for most images, quantified by Jaccard indices at around 90%. For images with smaller overlap of the segmentations, we demonstrate that this error is comparable to the variations across segmentations from different biological experts, suggesting that these images represent biologically unclear or ambiguous cases.
Authors: Ahmad Khalil, Tizian Dege, Pegah Golchin, Rostyslav Olshevskyi, Antonio Fernandez Anta, Tobias Meuser
Abstract: In the pursuit of refining precise perception models for fully autonomous driving, continual online model training becomes essential. Federated Learning (FL) within vehicular networks offers an efficient mechanism for model training while preserving raw sensory data integrity. Yet, FL struggles with non-identically distributed data (e.g., quantity skew), leading to suboptimal convergence rates during model training. In previous work, we introduced FedLA, an innovative Label-Aware aggregation method addressing data heterogeneity in FL for generic scenarios. In this paper, we introduce FedProx+LA, a novel FL method building upon the state-of-the-art FedProx and FedLA to tackle data heterogeneity, which is specifically tailored for vehicular networks. We evaluate the efficacy of FedProx+LA in continuous online object detection model training. Through a comparative analysis against conventional and state-of-the-art methods, our findings reveal the superior convergence rate of FedProx+LA. Notably, if the label distribution is very heterogeneous, our FedProx+LA approach shows substantial improvements in detection performance compared to baseline methods, also outperforming our previous FedLA approach. Moreover, both FedLA and FedProx+LA increase convergence speed by 30% compared to baseline methods.
Authors: Wenxuan Guo, Zhiyu Pan, Ziheng Xi, Alapati Tuerxun, Jianjiang Feng, Jie Zhou
Abstract: Sports analysis and viewing play a pivotal role in the current sports domain, offering significant value not only to coaches and athletes but also to fans and the media. In recent years, the rapid development of virtual reality (VR) and augmented reality (AR) technologies have introduced a new platform for watching games. Visualization of sports competitions in VR/AR represents a revolutionary technology, providing audiences with a novel immersive viewing experience. However, there is still a lack of related research in this area. In this work, we present for the first time a comprehensive system for sports competition analysis and real-time visualization on VR/AR platforms. First, we utilize multiview LiDARs and cameras to collect multimodal game data. Subsequently, we propose a framework for multi-player tracking and pose estimation based on a limited amount of supervised data, which extracts precise player positions and movements from point clouds and images. Moreover, we perform avatar modeling of players to obtain their 3D models. Ultimately, using these 3D player data, we conduct competition analysis and real-time visualization on VR/AR. Extensive quantitative experiments demonstrate the accuracy and robustness of our multi-player tracking and pose estimation framework. The visualization results showcase the immense potential of our sports visualization system on the domain of watching games on VR/AR devices. The multimodal competition dataset we collected and all related code will be released soon.
Authors: Seungyeop Lee, Knut Peterson, Solmaz Arezoomandan, Bill Cai, Peihan Li, Lifeng Zhou, David Han
Abstract: A major obstacle to the development of effective monocular depth estimation algorithms is the difficulty in obtaining high-quality depth data that corresponds to collected RGB images. Collecting this data is time-consuming and costly, and even data collected by modern sensors has limited range or resolution, and is subject to inconsistencies and noise. To combat this, we propose a method of data generation in simulation using 3D synthetic environments and CycleGAN domain transfer. We compare this method of data generation to the popular NYUDepth V2 dataset by training a depth estimation model based on the DenseDepth structure using different training sets of real and simulated data. We evaluate the performance of the models on newly collected images and LiDAR depth data from a Husky robot to verify the generalizability of the approach and show that GAN-transformed data can serve as an effective alternative to real-world data, particularly in depth estimation.
Authors: Alexander Geiger, Lars Wagner, Daniel Rueckert, Dirk Wilhelm, Alissa Jell
Abstract: High-resolution manometry (HRM) is the gold standard in diagnosing esophageal motility disorders. As HRM is typically conducted under short-term laboratory settings, intermittently occurring disorders are likely to be missed. Therefore, long-term (up to 24h) HRM (LTHRM) is used to gain detailed insights into the swallowing behavior. However, analyzing the extensive data from LTHRM is challenging and time consuming as medical experts have to analyze the data manually, which is slow and prone to errors. To address this challenge, we propose a Deep Learning based swallowing detection method to accurately identify swallowing events and secondary non-deglutitive-induced esophageal motility disorders in LTHRM data. We then proceed with clustering the identified swallows into distinct classes, which are analyzed by highly experienced clinicians to validate the different swallowing patterns. We evaluate our computational pipeline on a total of 25 LTHRMs, which were meticulously annotated by medical experts. By detecting more than 94% of all relevant swallow events and providing all relevant clusters for a more reliable diagnostic process among experienced clinicians, we are able to demonstrate the effectiveness as well as positive clinical impact of our approach to make LTHRM feasible in clinical care.
Authors: Mohammad Mahmudul Alam, Negin Sokhandan, Emmett Goodman
Abstract: In Virtual Product Placement (VPP) applications, the discrete integration of specific brand products into images or videos has emerged as a challenging yet important task. This paper introduces a novel three-stage fully automated VPP system. In the first stage, a language-guided image segmentation model identifies optimal regions within images for product inpainting. In the second stage, Stable Diffusion (SD), fine-tuned with a few example product images, is used to inpaint the product into the previously identified candidate regions. The final stage introduces an "Alignment Module", which is designed to effectively sieve out low-quality images. Comprehensive experiments demonstrate that the Alignment Module ensures the presence of the intended product in every generated image and enhances the average quality of images by 35%. The results presented in this paper demonstrate the effectiveness of the proposed VPP system, which holds significant potential for transforming the landscape of virtual advertising and marketing strategies.
Authors: Saahil Islam, Venkatesh N. Murthy, Dominik Neumann, Badhan Kumar Das, Puneet Sharma, Andreas Maier, Dorin Comaniciu, Florin C. Ghesu
Abstract: An accurate detection and tracking of devices such as guiding catheters in live X-ray image acquisitions is an essential prerequisite for endovascular cardiac interventions. This information is leveraged for procedural guidance, e.g., directing stent placements. To ensure procedural safety and efficacy, there is a need for high robustness no failures during tracking. To achieve that, one needs to efficiently tackle challenges, such as: device obscuration by contrast agent or other external devices or wires, changes in field-of-view or acquisition angle, as well as the continuous movement due to cardiac and respiratory motion. To overcome the aforementioned challenges, we propose a novel approach to learn spatio-temporal features from a very large data cohort of over 16 million interventional X-ray frames using self-supervision for image sequence data. Our approach is based on a masked image modeling technique that leverages frame interpolation based reconstruction to learn fine inter-frame temporal correspondences. The features encoded in the resulting model are fine-tuned downstream. Our approach achieves state-of-the-art performance and in particular robustness compared to ultra optimized reference solutions (that use multi-stage feature fusion, multi-task and flow regularization). The experiments show that our method achieves 66.31% reduction in maximum tracking error against reference solutions (23.20% when flow regularization is used); achieving a success score of 97.95% at a 3x faster inference speed of 42 frames-per-second (on GPU). The results encourage the use of our approach in various other tasks within interventional image analytics that require effective understanding of spatio-temporal semantics.
Authors: Daxin Li, Yuanchao Bai, Kai Wang, Junjun Jiang, Xianming Liu, Wen Gao
Abstract: Transformer-based entropy models have gained prominence in recent years due to their superior ability to capture long-range dependencies in probability distribution estimation compared to convolution-based methods. However, previous transformer-based entropy models suffer from a sluggish coding process due to pixel-wise autoregression or duplicated computation during inference. In this paper, we propose a novel transformer-based entropy model called GroupedMixer, which enjoys both faster coding speed and better compression performance than previous transformer-based methods. Specifically, our approach builds upon group-wise autoregression by first partitioning the latent variables into groups along spatial-channel dimensions, and then entropy coding the groups with the proposed transformer-based entropy model. The global causal self-attention is decomposed into more efficient group-wise interactions, implemented using inner-group and cross-group token-mixers. The inner-group token-mixer incorporates contextual elements within a group while the cross-group token-mixer interacts with previously decoded groups. Alternate arrangement of two token-mixers enables global contextual reference. To further expedite the network inference, we introduce context cache optimization to GroupedMixer, which caches attention activation values in cross-group token-mixers and avoids complex and duplicated computation. Experimental results demonstrate that the proposed GroupedMixer yields the state-of-the-art rate-distortion performance with fast compression speed.
Authors: Zijia Wang, Wenbin Yang, Zhisong Liu, Zhen Jia
Abstract: Self-training is a powerful approach to deep learning. The key process is to find a pseudo-label for modeling. However, previous self-training algorithms suffer from the over-confidence issue brought by the hard labels, even some confidence-related regularizers cannot comprehensively catch the uncertainty. Therefore, we propose a new self-training framework to combine uncertainty information of both model and dataset. Specifically, we propose to use Expectation-Maximization (EM) to smooth the labels and comprehensively estimate the uncertainty information. We further design a basis extraction network to estimate the initial basis from the dataset. The obtained basis with uncertainty can be filtered based on uncertainty information. It can then be transformed into the real hard label to iteratively update the model and basis in the retraining process. Experiments on image classification and semantic segmentation show the advantages of our methods among confidence-aware self-training algorithms with 1-3 percentage improvement on different datasets.
Authors: Zhiyu Pan, Yongjie Duan, Xiongjun Guan, Jianjiang Feng, Jie Zhou
Abstract: Latent fingerprint matching is a daunting task, primarily due to the poor quality of latent fingerprints. In this study, we propose a deep-learning based dense minutia descriptor (DMD) for latent fingerprint matching. A DMD is obtained by extracting the fingerprint patch aligned by its central minutia, capturing detailed minutia information and texture information. Our dense descriptor takes the form of a three-dimensional representation, with two dimensions associated with the original image plane and the other dimension representing the abstract features. Additionally, the extraction process outputs the fingerprint segmentation map, ensuring that the descriptor is only valid in the foreground region. The matching between two descriptors occurs in their overlapping regions, with a score normalization strategy to reduce the impact brought by the differences outside the valid area. Our descriptor achieves state-of-the-art performance on several latent fingerprint datasets. Overall, our DMD is more representative and interpretable compared to previous methods.
Authors: Yu Zhou, Xiahao Zou, Yi Wang
Abstract: Bone segmentation is an essential step for the preoperative planning of fracture trauma surgery. The automated segmentation of fractured bone from computed tomography (CT) scans remains challenging, due to the large differences of fractures in position and morphology, and also the inherent anatomical characteristics of different bone structures. To alleviate these issues, we propose a cross-scale attention mechanism as well as a surface supervision strategy for fractured bone segmentation in CT. Specifically, a cross-scale attention mechanism is introduced to effectively aggregate the features among different scales to provide more powerful fracture representation. Moreover, a surface supervision strategy is employed, which explicitly constrains the network to pay more attention to the bone boundary. The efficacy of the proposed method is evaluated on a public dataset containing CT scans with hip fractures. The evaluation metrics are Dice similarity coefficient (DSC), average symmetric surface distance (ASSD), and Hausdorff distance (95HD). The proposed method achieves an average DSC of 93.36%, ASSD of 0.85mm, 95HD of 7.51mm. Our method offers an effective fracture segmentation approach for the pelvic CT examinations, and has the potential to be used for improving the segmentation performance of other types of fractures.
Authors: Chenying Liu, Conrad Albrecht, Yi Wang, Xiao Xiang Zhu
Abstract: We study the potential of noisy labels y to pretrain semantic segmentation models in a multi-modal learning framework for geospatial applications. Specifically, we propose a novel Cross-modal Sample Selection method (CromSS) that utilizes the class distributions P^{(d)}(x,c) over pixels x and classes c modelled by multiple sensors/modalities d of a given geospatial scene. Consistency of predictions across sensors $d$ is jointly informed by the entropy of P^{(d)}(x,c). Noisy label sampling we determine by the confidence of each sensor d in the noisy class label, P^{(d)}(x,c=y(x)). To verify the performance of our approach, we conduct experiments with Sentinel-1 (radar) and Sentinel-2 (optical) satellite imagery from the globally-sampled SSL4EO-S12 dataset. We pair those scenes with 9-class noisy labels sourced from the Google Dynamic World project for pretraining. Transfer learning evaluations (downstream task) on the DFC2020 dataset confirm the effectiveness of the proposed method for remote sensing image segmentation.
Authors: Heng Li, Haojin Li, Jianyu Chen, Zhongxi Qiu, Huazhu Fu, Lidai Wang, Yan Hu, Jiang Liu
Abstract: Deep learning models often encounter challenges in making accurate inferences when there are domain shifts between the source and target data. This issue is particularly pronounced in clinical settings due to the scarcity of annotated data resulting from the professional and private nature of medical data. Despite the existence of decent solutions, many of them are hindered in clinical settings due to limitations in data collection and computational complexity. To tackle domain shifts in data-scarce medical scenarios, we propose a Random frequency filtering enabled Single-source Domain Generalization algorithm (RaffeSDG), which promises robust out-of-domain inference with segmentation models trained on a single-source domain. A filter-based data augmentation strategy is first proposed to promote domain variability within a single-source domain by introducing variations in frequency space and blending homologous samples. Then Gaussian filter-based structural saliency is also leveraged to learn robust representations across augmented samples, further facilitating the training of generalizable segmentation models. To validate the effectiveness of RaffeSDG, we conducted extensive experiments involving out-of-domain inference on segmentation tasks for three human tissues imaged by four diverse modalities. Through thorough investigations and comparisons, compelling evidence was observed in these experiments, demonstrating the potential and generalizability of RaffeSDG. The code is available at https://github.com/liamheng/Non-IID_Medical_Image_Segmentation.
URLs: https://github.com/liamheng/Non-IID_Medical_Image_Segmentation.
Authors: Nhi Nguyen, Le Nguyen, Honghan Li, Miguel Bordallo L\'opez, Constantino \'Alvarez Casado
Abstract: Video-based remote photoplethysmography (rPPG) has emerged as a promising technology for non-contact vital sign monitoring, especially under controlled conditions. However, the accurate measurement of vital signs in real-world scenarios faces several challenges, including artifacts induced by videocodecs, low-light noise, degradation, low dynamic range, occlusions, and hardware and network constraints. In this article, we systematically investigate comprehensive investigate these issues, measuring their detrimental effects on the quality of rPPG measurements. Additionally, we propose practical strategies for mitigating these challenges to improve the dependability and resilience of video-based rPPG systems. We detail methods for effective biosignal recovery in the presence of network limitations and present denoising and inpainting techniques aimed at preserving video frame integrity. Through extensive evaluations and direct comparisons, we demonstrate the effectiveness of the approaches in enhancing rPPG measurements under challenging environments, contributing to the development of more reliable and effective remote vital sign monitoring technologies.
Authors: Kai Luo, Hao Wu, Kefu Yi, Kailun Yang, Wei Hao, Rongdong Hu
Abstract: As human-machine interaction continues to evolve, the capacity for environmental perception is becoming increasingly crucial. Integrating the two most common types of sensory data, images, and point clouds, can enhance detection accuracy. However, currently, no model exists that can simultaneously detect an object's position in both point clouds and images and ascertain their corresponding relationship. This information is invaluable for human-machine interactions, offering new possibilities for their enhancement. In light of this, this paper introduces an end-to-end Consistency Object Detection (COD) algorithm framework that requires only a single forward inference to simultaneously obtain an object's position in both point clouds and images and establish their correlation. Furthermore, to assess the accuracy of the object correlation between point clouds and images, this paper proposes a new evaluation metric, Consistency Precision (CP). To verify the effectiveness of the proposed framework, an extensive set of experiments has been conducted on the KITTI and DAIR-V2X datasets. The study also explored how the proposed consistency detection method performs on images when the calibration parameters between images and point clouds are disturbed, compared to existing post-processing methods. The experimental results demonstrate that the proposed method exhibits excellent detection performance and robustness, achieving end-to-end consistency detection. The source code will be made publicly available at https://github.com/xifen523/COD.
Authors: Praveen Kumar Chandaliya, Kiran Raja, Raghavendra Ramachandra, Zahid Akhtar, Christoph Busch
Abstract: Numerous studies have shown that existing Face Recognition Systems (FRS), including commercial ones, often exhibit biases toward certain ethnicities due to under-represented data. In this work, we explore ethnicity alteration and skin tone modification using synthetic face image generation methods to increase the diversity of datasets. We conduct a detailed analysis by first constructing a balanced face image dataset representing three ethnicities: Asian, Black, and Indian. We then make use of existing Generative Adversarial Network-based (GAN) image-to-image translation and manifold learning models to alter the ethnicity from one to another. A systematic analysis is further conducted to assess the suitability of such datasets for FRS by studying the realistic skin-tone representation using Individual Typology Angle (ITA). Further, we also analyze the quality characteristics using existing Face image quality assessment (FIQA) approaches. We then provide a holistic FRS performance analysis using four different systems. Our findings pave the way for future research works in (i) developing both specific ethnicity and general (any to any) ethnicity alteration models, (ii) expanding such approaches to create databases with diverse skin tones, (iii) creating datasets representing various ethnicities which further can help in mitigating bias while addressing privacy concerns.
Authors: Shanshan Zhang, Mingqian Ji, Yang Li, Jian Yang
Abstract: Pedestrian detection has significantly progressed in recent years, thanks to the development of DNNs. However, detection performance at occluded scenes is still far from satisfactory, as occlusion increases the intra-class variance of pedestrians, hindering the model from finding an accurate classification boundary between pedestrians and background clutters. From the perspective of reducing intra-class variance, we propose to complete features for occluded regions so as to align the features of pedestrians across different occlusion patterns. An important premise for feature completion is to locate occluded regions. From our analysis, channel features of different pedestrian proposals only show high correlation values at visible parts and thus feature correlations can be used to model occlusion patterns. In order to narrow down the gap between completed features and real fully visible ones, we propose an adversarial learning method, which completes occluded features with a generator such that they can hardly be distinguished by the discriminator from real fully visible features. We report experimental results on the CityPersons, Caltech and CrowdHuman datasets. On CityPersons, we show significant improvements over five different baseline detectors, especially on the heavy occlusion subset. Furthermore, we show that our proposed method FeatComp++ achieves state-of-the-art results on all the above three datasets without relying on extra cues.
Authors: Zhiwei Xiong, Yunfan Zhang, Zhiqi Shen, Peiran Ren, Han Yu
Abstract: Image aesthetics assessment (IAA) is attracting wide interest with the prevalence of social media. The problem is challenging due to its subjective and ambiguous nature. Instead of directly extracting aesthetic features solely from the image, user comments associated with an image could potentially provide complementary knowledge that is useful for IAA. With existing large-scale pre-trained models demonstrating strong capabilities in extracting high-quality transferable visual and textual features, learnable queries are shown to be effective in extracting useful features from the pre-trained visual features. Therefore, in this paper, we propose MMLQ, which utilizes multi-modal learnable queries to extract aesthetics-related features from multi-modal pre-trained features. Extensive experimental results demonstrate that MMLQ achieves new state-of-the-art performance on multi-modal IAA, beating previous methods by 7.7% and 8.3% in terms of SRCC and PLCC, respectively.
Authors: Hoang-Quan Nguyen, Thanh-Dat Truong, Khoa Luu
Abstract: Action recognition has become one of the popular research topics in computer vision. There are various methods based on Convolutional Networks and self-attention mechanisms as Transformers to solve both spatial and temporal dimensions problems of action recognition tasks that achieve competitive performances. However, these methods lack a guarantee of the correctness of the action subject that the models give attention to, i.e., how to ensure an action recognition model focuses on the proper action subject to make a reasonable action prediction. In this paper, we propose a multi-view attention consistency method that computes the similarity between two attentions from two different views of the action videos using Directed Gromov-Wasserstein Discrepancy. Furthermore, our approach applies the idea of Neural Radiance Field to implicitly render the features from novel views when training on single-view datasets. Therefore, the contributions in this work are three-fold. Firstly, we introduce the multi-view attention consistency to solve the problem of reasonable prediction in action recognition. Secondly, we define a new metric for multi-view consistent attention using Directed Gromov-Wasserstein Discrepancy. Thirdly, we built an action recognition model based on Video Transformers and Neural Radiance Fields. Compared to the recent action recognition methods, the proposed approach achieves state-of-the-art results on three large-scale datasets, i.e., Jester, Something-Something V2, and Kinetics-400.
Authors: Yik Lung Pang, Changjae Oh, Andrea Cavallaro
Abstract: Recent works in hand-object reconstruction mainly focus on the single-view and dense multi-view settings. On the one hand, single-view methods can leverage learned shape priors to generalise to unseen objects but are prone to inaccuracies due to occlusions. On the other hand, dense multi-view methods are very accurate but cannot easily adapt to unseen objects without further data collection. In contrast, sparse multi-view methods can take advantage of the additional views to tackle occlusion, while keeping the computational cost low compared to dense multi-view methods. In this paper, we consider the problem of hand-object reconstruction with unseen objects in the sparse multi-view setting. Given multiple RGB images of the hand and object captured at the same time, our model SVHO combines the predictions from each view into a unified reconstruction without optimisation across views. We train our model on a synthetic hand-object dataset and evaluate directly on a real world recorded hand-object dataset with unseen objects. We show that while reconstruction of unseen hands and objects from RGB is challenging, additional views can help improve the reconstruction quality.
Authors: Kelvin C. K. Chan, Yang Zhao, Xuhui Jia, Ming-Hsuan Yang, Huisheng Wang
Abstract: In subject-driven text-to-image synthesis, the synthesis process tends to be heavily influenced by the reference images provided by users, often overlooking crucial attributes detailed in the text prompt. In this work, we propose Subject-Agnostic Guidance (SAG), a simple yet effective solution to remedy the problem. We show that through constructing a subject-agnostic condition and applying our proposed dual classifier-free guidance, one could obtain outputs consistent with both the given subject and input text prompts. We validate the efficacy of our approach through both optimization-based and encoder-based methods. Additionally, we demonstrate its applicability in second-order customization methods, where an encoder-based model is fine-tuned with DreamBooth. Our approach is conceptually simple and requires only minimal code modifications, but leads to substantial quality improvements, as evidenced by our evaluations and user studies.
Authors: Samir Khaki, Ahmad Sajedi, Kai Wang, Lucy Z. Liu, Yuri A. Lawryshyn, Konstantinos N. Plataniotis
Abstract: Recent works in dataset distillation seek to minimize training expenses by generating a condensed synthetic dataset that encapsulates the information present in a larger real dataset. These approaches ultimately aim to attain test accuracy levels akin to those achieved by models trained on the entirety of the original dataset. Previous studies in feature and distribution matching have achieved significant results without incurring the costs of bi-level optimization in the distillation process. Despite their convincing efficiency, many of these methods suffer from marginal downstream performance improvements, limited distillation of contextual information, and subpar cross-architecture generalization. To address these challenges in dataset distillation, we propose the ATtentiOn Mixer (ATOM) module to efficiently distill large datasets using a mixture of channel and spatial-wise attention in the feature matching process. Spatial-wise attention helps guide the learning process based on consistent localization of classes in their respective images, allowing for distillation from a broader receptive field. Meanwhile, channel-wise attention captures the contextual information associated with the class itself, thus making the synthetic image more informative for training. By integrating both types of attention, our ATOM module demonstrates superior performance across various computer vision datasets, including CIFAR10/100 and TinyImagenet. Notably, our method significantly improves performance in scenarios with a low number of images per class, thereby enhancing its potential. Furthermore, we maintain the improvement in cross-architectures and applications such as neural architecture search.
Authors: Abdoul Aziz Amadou, Vivek Singh, Florin C. Ghesu, Young-Ho Kim, Laura Stanciulescu, Harshitha P. Sai, Puneet Sharma, Alistair Young, Ronak Rajani, Kawal Rhode
Abstract: Transesophageal echocardiography (TEE) plays a pivotal role in cardiology for diagnostic and interventional procedures. However, using it effectively requires extensive training due to the intricate nature of image acquisition and interpretation. To enhance the efficiency of novice sonographers and reduce variability in scan acquisitions, we propose a novel ultrasound (US) navigation assistance method based on contrastive learning as goal-conditioned reinforcement learning (GCRL). We augment the previous framework using a novel contrastive patient batching method (CPB) and a data-augmented contrastive loss, both of which we demonstrate are essential to ensure generalization to anatomical variations across patients. The proposed framework enables navigation to both standard diagnostic as well as intricate interventional views with a single model. Our method was developed with a large dataset of 789 patients and obtained an average error of 6.56 mm in position and 9.36 degrees in angle on a testing dataset of 140 patients, which is competitive or superior to models trained on individual views. Furthermore, we quantitatively validate our method's ability to navigate to interventional views such as the Left Atrial Appendage (LAA) view used in LAA closure. Our approach holds promise in providing valuable guidance during transesophageal ultrasound examinations, contributing to the advancement of skill acquisition for cardiac ultrasound practitioners.
Authors: Yuan Tang, Xu Han, Xianzhi Li, Qiao Yu, Yixue Hao, Long Hu, Min Chen
Abstract: Large 2D vision-language models (2D-LLMs) have gained significant attention by bridging Large Language Models (LLMs) with images using a simple projector. Inspired by their success, large 3D point cloud-language models (3D-LLMs) also integrate point clouds into LLMs. However, directly aligning point clouds with LLM requires expensive training costs, typically in hundreds of GPU-hours on A100, which hinders the development of 3D-LLMs. In this paper, we introduce MiniGPT-3D, an efficient and powerful 3D-LLM that achieves multiple SOTA results while training for only 27 hours on one RTX 3090. Specifically, we propose to align 3D point clouds with LLMs using 2D priors from 2D-LLMs, which can leverage the similarity between 2D and 3D visual information. We introduce a novel four-stage training strategy for modality alignment in a cascaded way, and a mixture of query experts module to adaptively aggregate features with high efficiency. Moreover, we utilize parameter-efficient fine-tuning methods LoRA and Norm fine-tuning, resulting in only 47.8M learnable parameters, which is up to 260x fewer than existing methods. Extensive experiments show that MiniGPT-3D achieves SOTA on 3D object classification and captioning tasks, with significantly cheaper training costs. Notably, MiniGPT-3D gains an 8.12 increase on GPT-4 evaluation score for the challenging object captioning task compared to ShapeLLM-13B, while the latter costs 160 total GPU-hours on 8 A800. We are the first to explore the efficient 3D-LLM, offering new insights to the community. Code and weights are available at https://github.com/TangYuan96/MiniGPT-3D.
Authors: Yupeng Zhou, Daquan Zhou, Ming-Ming Cheng, Jiashi Feng, Qibin Hou
Abstract: For recent diffusion-based generative models, maintaining consistent content across a series of generated images, especially those containing subjects and complex details, presents a significant challenge. In this paper, we propose a new way of self-attention calculation, termed Consistent Self-Attention, that significantly boosts the consistency between the generated images and augments prevalent pretrained diffusion-based text-to-image models in a zero-shot manner. To extend our method to long-range video generation, we further introduce a novel semantic space temporal motion prediction module, named Semantic Motion Predictor. It is trained to estimate the motion conditions between two provided images in the semantic spaces. This module converts the generated sequence of images into videos with smooth transitions and consistent subjects that are significantly more stable than the modules based on latent spaces only, especially in the context of long video generation. By merging these two novel components, our framework, referred to as StoryDiffusion, can describe a text-based story with consistent images or videos encompassing a rich variety of contents. The proposed StoryDiffusion encompasses pioneering explorations in visual story generation with the presentation of images and videos, which we hope could inspire more research from the aspect of architectural modifications. Our code is made publicly available at https://github.com/HVision-NKU/StoryDiffusion.
Authors: Ruijie Zhao, Pinyan Tang, Sihui Luo
Abstract: Despite remarkable advancements, mainstream gaze estimation techniques, particularly appearance-based methods, often suffer from performance degradation in uncontrolled environments due to variations in illumination and individual facial attributes. Existing domain adaptation strategies, limited by their need for target domain samples, may fall short in real-world applications. This letter introduces Branch-out Auxiliary Regularization (BAR), an innovative method designed to boost gaze estimation's generalization capabilities without requiring direct access to target domain data. Specifically, BAR integrates two auxiliary consistency regularization branches: one that uses augmented samples to counteract environmental variations, and another that aligns gaze directions with positive source domain samples to encourage the learning of consistent gaze features. These auxiliary pathways strengthen the core network and are integrated in a smooth, plug-and-play manner, facilitating easy adaptation to various other models. Comprehensive experimental evaluations on four cross-dataset tasks demonstrate the superiority of our approach.
Authors: Wenshuo Chen, Hongru Xiao, Erhang Zhang, Lijie Hu, Lei Wang, Mengyuan Liu, Chen Chen
Abstract: Is the Text to Motion model robust? Recent advancements in Text to Motion models primarily stem from more accurate predictions of specific actions. However, the text modality typically relies solely on pre-trained Contrastive Language-Image Pretraining (CLIP) models. Our research has uncovered a significant issue with the text-to-motion model: its predictions often exhibit inconsistent outputs, resulting in vastly different or even incorrect poses when presented with semantically similar or identical text inputs. In this paper, we undertake an analysis to elucidate the underlying causes of this instability, establishing a clear link between the unpredictability of model outputs and the erratic attention patterns of the text encoder module. Consequently, we introduce a formal framework aimed at addressing this issue, which we term the Stable Text-to-Motion Framework (SATO). SATO consists of three modules, each dedicated to stable attention, stable prediction, and maintaining a balance between accuracy and robustness trade-off. We present a methodology for constructing an SATO that satisfies the stability of attention and prediction. To verify the stability of the model, we introduced a new textual synonym perturbation dataset based on HumanML3D and KIT-ML. Results show that SATO is significantly more stable against synonyms and other slight perturbations while keeping its high accuracy performance.
Authors: Th\'eo Moutakanni, Piotr Bojanowski, Guillaume Chassagnon, C\'eline Hudelot, Armand Joulin, Yann LeCun, Matthew Muckley, Maxime Oquab, Marie-Pierre Revel, Maria Vakalopoulou
Abstract: AI Foundation models are gaining traction in various applications, including medical fields like radiology. However, medical foundation models are often tested on limited tasks, leaving their generalisability and biases unexplored. We present RayDINO, a large visual encoder trained by self-supervision on 873k chest X-rays. We compare RayDINO to previous state-of-the-art models across nine radiology tasks, from classification and dense segmentation to text generation, and provide an in depth analysis of population, age and sex biases of our model. Our findings suggest that self-supervision allows patient-centric AI proving useful in clinical workflows and interpreting X-rays holistically. With RayDINO and small task-specific adapters, we reach state-of-the-art results and improve generalization to unseen populations while mitigating bias, illustrating the true promise of foundation models: versatility and robustness.
Authors: Dongfu Jiang, Xuan He, Huaye Zeng, Cong Wei, Max Ku, Qian Liu, Wenhu Chen
Abstract: The recent years have witnessed a great array of large multimodal models (LMMs) to effectively solve single-image vision language tasks. However, their abilities to solve multi-image visual language tasks is yet to be improved. The existing multi-image LMMs (e.g. OpenFlamingo, Emu, Idefics, etc) mostly gain their multi-image ability through pre-training on hundreds of millions of noisy interleaved image-text data from web, which is neither efficient nor effective. In this paper, we aim at building strong multi-image LMMs via instruction tuning with academic-level resources. Therefore, we meticulously construct Mantis-Instruct containing 721K instances from 14 multi-image datasets. We design Mantis-Instruct to cover different multi-image skills like co-reference, reasoning, comparing, temporal understanding. We combine Mantis-Instruct with several single-image visual-language datasets to train our model Mantis to handle any interleaved image-text inputs. We evaluate the trained Mantis on five multi-image benchmarks and eight single-image benchmarks. Though only requiring academic-level resources (i.e. 36 hours on 16xA100-40G), Mantis-8B can achieve state-of-the-art performance on all the multi-image benchmarks and beats the existing best multi-image LMM Idefics2-8B by an average of 9 absolute points. We observe that Mantis performs equivalently well on the held-in and held-out evaluation benchmarks. We further evaluate Mantis on single-image benchmarks and demonstrate that Mantis can maintain a strong single-image performance on par with CogVLM and Emu2. Our results are particularly encouraging as it shows that low-cost instruction tuning is indeed much more effective than intensive pre-training in terms of building multi-image LMMs.
Authors: Matias Mendieta, Guangyu Sun, Chen Chen
Abstract: Federated learning (FL) enables multiple clients to train models collectively while preserving data privacy. However, FL faces challenges in terms of communication cost and data heterogeneity. One-shot federated learning has emerged as a solution by reducing communication rounds, improving efficiency, and providing better security against eavesdropping attacks. Nevertheless, data heterogeneity remains a significant challenge, impacting performance. This work explores the effectiveness of diffusion models in one-shot FL, demonstrating their applicability in addressing data heterogeneity and improving FL performance. Additionally, we investigate the utility of our diffusion model approach, FedDiff, compared to other one-shot FL methods under differential privacy (DP). Furthermore, to improve generated sample quality under DP settings, we propose a pragmatic Fourier Magnitude Filtering (FMF) method, enhancing the effectiveness of generated data for global model training.
Authors: Chuanming Tang, Kai Wang, Fei Yang, Joost van de Weijer
Abstract: Large-scale Text-to-Image (T2I) diffusion models demonstrate significant generation capabilities based on textual prompts. Based on the T2I diffusion models, text-guided image editing research aims to empower users to manipulate generated images by altering the text prompts. However, existing image editing techniques are prone to editing over unintentional regions that are beyond the intended target area, primarily due to inaccuracies in cross-attention maps. To address this problem, we propose Localization-aware Inversion (LocInv), which exploits segmentation maps or bounding boxes as extra localization priors to refine the cross-attention maps in the denoising phases of the diffusion process. Through the dynamic updating of tokens corresponding to noun words in the textual input, we are compelling the cross-attention maps to closely align with the correct noun and adjective words in the text prompt. Based on this technique, we achieve fine-grained image editing over particular objects while preventing undesired changes to other regions. Our method LocInv, based on the publicly available Stable Diffusion, is extensively evaluated on a subset of the COCO dataset, and consistently obtains superior results both quantitatively and qualitatively.The code will be released at https://github.com/wangkai930418/DPL
Authors: Matin Mortaheb, Erciyes Karakaya, Mohammad A. Amir Khojastepour, Sennur Ulukus
Abstract: The transformer structure employed in large language models (LLMs), as a specialized category of deep neural networks (DNNs) featuring attention mechanisms, stands out for their ability to identify and highlight the most relevant aspects of input data. Such a capability is particularly beneficial in addressing a variety of communication challenges, notably in the realm of semantic communication where proper encoding of the relevant data is critical especially in systems with limited bandwidth. In this work, we employ vision transformers specifically for the purpose of compression and compact representation of the input image, with the goal of preserving semantic information throughout the transmission process. Through the use of the attention mechanism inherent in transformers, we create an attention mask. This mask effectively prioritizes critical segments of images for transmission, ensuring that the reconstruction phase focuses on key objects highlighted by the mask. Our methodology significantly improves the quality of semantic communication and optimizes bandwidth usage by encoding different parts of the data in accordance with their semantic information content, thus enhancing overall efficiency. We evaluate the effectiveness of our proposed framework using the TinyImageNet dataset, focusing on both reconstruction quality and accuracy. Our evaluation results demonstrate that our framework successfully preserves semantic information, even when only a fraction of the encoded data is transmitted, according to the intended compression rates.
Authors: Shihao Wang, Zhiding Yu, Xiaohui Jiang, Shiyi Lan, Min Shi, Nadine Chang, Jan Kautz, Ying Li, Jose M. Alvarez
Abstract: The advances in multimodal large language models (MLLMs) have led to growing interests in LLM-based autonomous driving agents to leverage their strong reasoning capabilities. However, capitalizing on MLLMs' strong reasoning capabilities for improved planning behavior is challenging since planning requires full 3D situational awareness beyond 2D reasoning. To address this challenge, our work proposes a holistic framework for strong alignment between agent models and 3D driving tasks. Our framework starts with a novel 3D MLLM architecture that uses sparse queries to lift and compress visual representations into 3D before feeding them into an LLM. This query-based representation allows us to jointly encode dynamic objects and static map elements (e.g., traffic lanes), providing a condensed world model for perception-action alignment in 3D. We further propose OmniDrive-nuScenes, a new visual question-answering dataset challenging the true 3D situational awareness of a model with comprehensive visual question-answering (VQA) tasks, including scene description, traffic regulation, 3D grounding, counterfactual reasoning, decision making and planning. Extensive studies show the effectiveness of the proposed architecture as well as the importance of the VQA tasks for reasoning and planning in complex 3D scenes.
Authors: Maxwell Jones, Sheng-Yu Wang, Nupur Kumari, David Bau, Jun-Yan Zhu
Abstract: Art reinterpretation is the practice of creating a variation of a reference work, making a paired artwork that exhibits a distinct artistic style. We ask if such an image pair can be used to customize a generative model to capture the demonstrated stylistic difference. We propose Pair Customization, a new customization method that learns stylistic difference from a single image pair and then applies the acquired style to the generation process. Unlike existing methods that learn to mimic a single concept from a collection of images, our method captures the stylistic difference between paired images. This allows us to apply a stylistic change without overfitting to the specific image content in the examples. To address this new task, we employ a joint optimization method that explicitly separates the style and content into distinct LoRA weight spaces. We optimize these style and content weights to reproduce the style and content images while encouraging their orthogonality. During inference, we modify the diffusion process via a new style guidance based on our learned weights. Both qualitative and quantitative experiments show that our method can effectively learn style while avoiding overfitting to image content, highlighting the potential of modeling such stylistic differences from a single image pair.
Authors: Youquan Liu, Lingdong Kong, Xiaoyang Wu, Runnan Chen, Xin Li, Liang Pan, Ziwei Liu, Yuexin Ma
Abstract: A unified and versatile LiDAR segmentation model with strong robustness and generalizability is desirable for safe autonomous driving perception. This work presents M3Net, a one-of-a-kind framework for fulfilling multi-task, multi-dataset, multi-modality LiDAR segmentation in a universal manner using just a single set of parameters. To better exploit data volume and diversity, we first combine large-scale driving datasets acquired by different types of sensors from diverse scenes and then conduct alignments in three spaces, namely data, feature, and label spaces, during the training. As a result, M3Net is capable of taming heterogeneous data for training state-of-the-art LiDAR segmentation models. Extensive experiments on twelve LiDAR segmentation datasets verify our effectiveness. Notably, using a shared set of parameters, M3Net achieves 75.1%, 83.1%, and 72.4% mIoU scores, respectively, on the official benchmarks of SemanticKITTI, nuScenes, and Waymo Open.
Authors: Anna Zapaishchykova, Benjamin H. Kann, Divyanshu Tak, Zezhong Ye, Daphne A. Haas-Kogan, Hugo J. W. L. Aerts
Abstract: Synthetic longitudinal brain MRI simulates brain aging and would enable more efficient research on neurodevelopmental and neurodegenerative conditions. Synthetically generated, age-adjusted brain images could serve as valuable alternatives to costly longitudinal imaging acquisitions, serve as internal controls for studies looking at the effects of environmental or therapeutic modifiers on brain development, and allow data augmentation for diverse populations. In this paper, we present a diffusion-based approach called SynthBrainGrow for synthetic brain aging with a two-year step. To validate the feasibility of using synthetically-generated data on downstream tasks, we compared structural volumetrics of two-year-aged brains against synthetically-aged brain MRI. Results show that SynthBrainGrow can accurately capture substructure volumetrics and simulate structural changes such as ventricle enlargement and cortical thinning. Our approach provides a novel way to generate longitudinal brain datasets from cross-sectional data to enable augmented training and benchmarking of computational tools for analyzing lifespan trajectories. This work signifies an important advance in generative modeling to synthesize realistic longitudinal data with limited lifelong MRI scans. The code is available at XXX.
Authors: ZhenZhou Wang
Abstract: The visual sensing system is one of the most important parts of the welding robots to realize intelligent and autonomous welding. The active visual sensing methods have been widely adopted in robotic welding because of their higher accuracies compared to the passive visual sensing methods. In this paper, we give a comprehensive review of the active visual sensing methods for robotic welding. According to their uses, we divide the state-of-the-art active visual sensing methods into four categories: seam tracking, weld bead defect detection, 3D weld pool geometry measurement and welding path planning. Firstly, we review the principles of these active visual sensing methods. Then, we give a tutorial of the 3D calibration methods for the active visual sensing systems used in intelligent welding robots to fill the gaps in the related fields. At last, we compare the reviewed active visual sensing methods and give the prospects based on their advantages and disadvantages.
Authors: Vadim Tynchenko, Aleksei Kudryavtsev, Vladimir Nelyub, Aleksei Borodulin, Andrei Gantimurov
Abstract: This report presents the test results Python library BaumEvA, which implements evolutionary algorithms for optimizing various types of problems, including computer vision tasks accompanied by the search for optimal model architectures. Testing was carried out to evaluate the effectiveness and reliability of the pro-posed methods, as well as to determine their applicability in various fields. Dur-ing testing, various test functions and parameters of evolutionary algorithms were used, which made it possible to evaluate their performance in a wide range of conditions. Test results showed that the library provides effective and reliable methods for solving optimization problems. However, some limitations were identified related to computational resources and execution time of algorithms on problems with large dimensions. The report includes a detailed description of the tests performed, the results obtained and conclusions about the applicability of the genetic algorithm in various tasks. Recommendations for choosing algorithm pa-rameters and using the library to achieve the best results are also provided. The report may be useful to developers involved in the optimization of complex com-puting systems, as well as to researchers studying the possibilities of using evo-lutionary algorithms in various fields of science and technology.
Authors: Chenqi Guo, Shiwei Zhong, Xiaofeng Liu, Qianli Feng, Yinglong Ma
Abstract: Does Knowledge Distillation (KD) really work? Conventional wisdom viewed it as a knowledge transfer procedure where a perfect mimicry of the student to its teacher is desired. However, paradoxical studies indicate that closely replicating the teacher's behavior does not consistently improve student generalization, posing questions on its possible causes. Confronted with this gap, we hypothesize that diverse attentions in teachers contribute to better student generalization at the expense of reduced fidelity in ensemble KD setups. By increasing data augmentation strengths, our key findings reveal a decrease in the Intersection over Union (IoU) of attentions between teacher models, leading to reduced student overfitting and decreased fidelity. We propose this low-fidelity phenomenon as an underlying characteristic rather than a pathology when training KD. This suggests that stronger data augmentation fosters a broader perspective provided by the divergent teacher ensemble and lower student-teacher mutual information, benefiting generalization performance. These insights clarify the mechanism on low-fidelity phenomenon in KD. Thus, we offer new perspectives on optimizing student model performance, by emphasizing increased diversity in teacher attentions and reduced mimicry behavior between teachers and student.
Authors: Jiahui Li, Tianle Shen, Zekai Gu, Jiawei Sun, Chengran Yuan, Yuhang Han, Shuo Sun, Marcelo H. Ang Jr
Abstract: Motion prediction is a challenging problem in autonomous driving as it demands the system to comprehend stochastic dynamics and the multi-modal nature of real-world agent interactions. Diffusion models have recently risen to prominence, and have proven particularly effective in pedestrian motion prediction tasks. However, the significant time consumption and sensitivity to noise have limited the real-time predictive capability of diffusion models. In response to these impediments, we propose a novel diffusion-based, acceleratable framework that adeptly predicts future trajectories of agents with enhanced resistance to noise. The core idea of our model is to learn a coarse-grained prior distribution of trajectory, which can skip a large number of denoise steps. This advancement not only boosts sampling efficiency but also maintains the fidelity of prediction accuracy. Our method meets the rigorous real-time operational standards essential for autonomous vehicles, enabling prompt trajectory generation that is vital for secure and efficient navigation. Through extensive experiments, our method speeds up the inference time to 136ms compared to standard diffusion model, and achieves significant improvement in multi-agent motion prediction on the Argoverse 1 motion forecasting dataset.
Authors: Zhenya Yang, Kai Chen, Yonghao Long, Qi Dou
Abstract: Surgical scene simulation plays a crucial role in surgical education and simulator-based robot learning. Traditional approaches for creating these environments with surgical scene involve a labor-intensive process where designers hand-craft tissues models with textures and geometries for soft body simulations. This manual approach is not only time-consuming but also limited in the scalability and realism. In contrast, data-driven simulation offers a compelling alternative. It has the potential to automatically reconstruct 3D surgical scenes from real-world surgical video data, followed by the application of soft body physics. This area, however, is relatively uncharted. In our research, we introduce 3D Gaussian as a learnable representation for surgical scene, which is learned from stereo endoscopic video. To prevent over-fitting and ensure the geometrical correctness of these scenes, we incorporate depth supervision and anisotropy regularization into the Gaussian learning process. Furthermore, we apply the Material Point Method, which is integrated with physical properties, to the 3D Gaussians to achieve realistic scene deformations. Our method was evaluated on our collected in-house and public surgical videos datasets. Results show that it can reconstruct and simulate surgical scenes from endoscopic videos efficiently-taking only a few minutes to reconstruct the surgical scene-and produce both visually and physically plausible deformations at a speed approaching real-time. The results demonstrate great potential of our proposed method to enhance the efficiency and variety of simulations available for surgical education and robot learning.
Authors: Zhe Niu, Ronglai Zuo, Brian Mak, Fangyun Wei
Abstract: This paper introduces TVB-HKSL-News, a new Hong Kong sign language (HKSL) dataset collected from a TV news program over a period of 7 months. The dataset is collected to enrich resources for HKSL and support research in large-vocabulary continuous sign language recognition (SLR) and translation (SLT). It consists of 16.07 hours of sign videos of two signers with a vocabulary of 6,515 glosses (for SLR) and 2,850 Chinese characters or 18K Chinese words (for SLT). One signer has 11.66 hours of sign videos and the other has 4.41 hours. One objective in building the dataset is to support the investigation of how well large-vocabulary continuous sign language recognition/translation can be done for a single signer given a (relatively) large amount of his/her training data, which could potentially lead to the development of new modeling methods. Besides, most parts of the data collection pipeline are automated with little human intervention; we believe that our collection method can be scaled up to collect more sign language data easily for SLT in the future for any sign languages if such sign-interpreted videos are available. We also run a SOTA SLR/SLT model on the dataset and get a baseline SLR word error rate of 34.08% and a baseline SLT BLEU-4 score of 23.58 for benchmarking future research on the dataset.
Authors: Yongxian Wei, Zixuan Hu, Zhenyi Wang, Li Shen, Chun Yuan, Dacheng Tao
Abstract: Data-Free Meta-Learning (DFML) aims to extract knowledge from a collection of pre-trained models without requiring the original data, presenting practical benefits in contexts constrained by data privacy concerns. Current DFML methods primarily focus on the data recovery from these pre-trained models. However, they suffer from slow recovery speed and overlook gaps inherent in heterogeneous pre-trained models. In response to these challenges, we introduce the Faster and Better Data-Free Meta-Learning (FREE) framework, which contains: (i) a meta-generator for rapidly recovering training tasks from pre-trained models; and (ii) a meta-learner for generalizing to new unseen tasks. Specifically, within the module Faster Inversion via Meta-Generator, each pre-trained model is perceived as a distinct task. The meta-generator can rapidly adapt to a specific task in just five steps, significantly accelerating the data recovery. Furthermore, we propose Better Generalization via Meta-Learner and introduce an implicit gradient alignment algorithm to optimize the meta-learner. This is achieved as aligned gradient directions alleviate potential conflicts among tasks from heterogeneous pre-trained models. Empirical experiments on multiple benchmarks affirm the superiority of our approach, marking a notable speed-up (20$\times$) and performance enhancement (1.42\% $\sim$ 4.78\%) in comparison to the state-of-the-art.
Authors: Aditya Chakravarty
Abstract: Recent transformer-based ASR models have achieved word-error rates (WER) below 4%, surpassing human annotator accuracy, yet they demand extensive server resources, contributing to significant carbon footprints. The traditional server-based architecture of ASR also presents privacy concerns, alongside reliability and latency issues due to network dependencies. In contrast, on-device (edge) ASR enhances privacy, boosts performance, and promotes sustainability by effectively balancing energy use and accuracy for specific applications. This study examines the effects of quantization, memory demands, and energy consumption on the performance of various ASR model inference on the NVIDIA Jetson Orin Nano. By analyzing WER and transcription speed across models using FP32, FP16, and INT8 quantization on clean and noisy datasets, we highlight the crucial trade-offs between accuracy, speeds, quantization, energy efficiency, and memory needs. We found that changing precision from fp32 to fp16 halves the energy consumption for audio transcription across different models, with minimal performance degradation. A larger model size and number of parameters neither guarantees better resilience to noise, nor predicts the energy consumption for a given transcription load. These, along with several other findings offer novel insights for optimizing ASR systems within energy- and memory-limited environments, crucial for the development of efficient on-device ASR solutions. The code and input data needed to reproduce the results in this article are open sourced are available on [https://github.com/zzadiues3338/ASR-energy-jetson].
Authors: Alex Murphy, Joel Zylberberg, Alona Fyshe
Abstract: Centred Kernel Alignment (CKA) has recently emerged as a popular metric to compare activations from biological and artificial neural networks (ANNs) in order to quantify the alignment between internal representations derived from stimuli sets (e.g. images, text, video) that are presented to both systems. In this paper we highlight issues that the community should take into account if using CKA as an alignment metric with neural data. Neural data are in the low-data high-dimensionality domain, which is one of the cases where (biased) CKA results in high similarity scores even for pairs of random matrices. Using fMRI and MEG data from the THINGS project, we show that if biased CKA is applied to representations of different sizes in the low-data high-dimensionality domain, they are not directly comparable due to biased CKA's sensitivity to differing feature-sample ratios and not stimuli-driven responses. This situation can arise both when comparing a pre-selected area of interest (e.g. ROI) to multiple ANN layers, as well as when determining to which ANN layer multiple regions of interest (ROIs) / sensor groups of different dimensionality are most similar. We show that biased CKA can be artificially driven to its maximum value when using independent random data of different sample-feature ratios. We further show that shuffling sample-feature pairs of real neural data does not drastically alter biased CKA similarity in comparison to unshuffled data, indicating an undesirable lack of sensitivity to stimuli-driven neural responses. Positive alignment of true stimuli-driven responses is only achieved by using debiased CKA. Lastly, we report findings that suggest biased CKA is sensitive to the inherent structure of neural data, only differing from shuffled data when debiased CKA detects stimuli-driven alignment.
Authors: Zhongzheng Qiao, Xuan Huy Pham, Savitha Ramasamy, Xudong Jiang, Erdal Kayacan, Andriy Sarabakha
Abstract: In autonomous and mobile robotics, a principal challenge is resilient real-time environmental perception, particularly in situations characterized by unknown and dynamic elements, as exemplified in the context of autonomous drone racing. This study introduces a perception technique for detecting drone racing gates under illumination variations, which is common during high-speed drone flights. The proposed technique relies upon a lightweight neural network backbone augmented with capabilities for continual learning. The envisaged approach amalgamates predictions of the gates' positional coordinates, distance, and orientation, encapsulating them into a cohesive pose tuple. A comprehensive number of tests serve to underscore the efficacy of this approach in confronting diverse and challenging scenarios, specifically those involving variable lighting conditions. The proposed methodology exhibits notable robustness in the face of illumination variations, thereby substantiating its effectiveness.
Authors: Tong Lei, Brian N. Bailey
Abstract: Simulating soil reflectance spectra is invaluable for soil-plant radiative modeling and training machine learning models, yet it is difficult as the intricate relationships between soil structure and its constituents. To address this, a fully data-driven soil optics generative model (SOGM) for simulation of soil reflectance spectra based on soil property inputs was developed. The model is trained on an extensive dataset comprising nearly 180,000 soil spectra-property pairs from 17 datasets. It generates soil reflectance spectra from text-based inputs describing soil properties and their values rather than only numerical values and labels in binary vector format. The generative model can simulate output spectra based on an incomplete set of input properties. SOGM is based on the denoising diffusion probabilistic model (DDPM). Two additional sub-models were also built to complement the SOGM: a spectral padding model that can fill in the gaps for spectra shorter than the full visible-near-infrared range (VIS-NIR; 400 to 2499 nm), and a wet soil spectra model that can estimate the effects of water content on soil reflectance spectra given the dry spectrum predicted by the SOGM. The SOGM was up-scaled by coupling with the Helios 3D plant modeling software, which allowed for generation of synthetic aerial images of simulated soil and plant scenes. It can also be easily integrated with soil-plant radiation model used for remote sensin research like PROSAIL. The testing results of the SOGM on new datasets that not included in model training proved that the model can generate reasonable soil reflectance spectra based on available property inputs. The presented models are openly accessible on: https://github.com/GEMINI-Breeding/SOGM_soil_spectra_simulation.
URLs: https://github.com/GEMINI-Breeding/SOGM_soil_spectra_simulation.
Authors: Sonakshi Garg, Hugo J\"onsson, Gustav Kalander, Axel Nilsson, Bhhaanu Pirange, Viktor Valadi, Johan \"Ostman
Abstract: Federated Learning (FL) is a decentralized learning paradigm, enabling parties to collaboratively train models while keeping their data confidential. Within autonomous driving, it brings the potential of reducing data storage costs, reducing bandwidth requirements, and to accelerate the learning. FL is, however, susceptible to poisoning attacks. In this paper, we introduce two novel poisoning attacks on FL tailored to regression tasks within autonomous driving: FLStealth and Off-Track Attack (OTA). FLStealth, an untargeted attack, aims at providing model updates that deteriorate the global model performance while appearing benign. OTA, on the other hand, is a targeted attack with the objective to change the global model's behavior when exposed to a certain trigger. We demonstrate the effectiveness of our attacks by conducting comprehensive experiments pertaining to the task of vehicle trajectory prediction. In particular, we show that, among five different untargeted attacks, FLStealth is the most successful at bypassing the considered defenses employed by the server. For OTA, we demonstrate the inability of common defense strategies to mitigate the attack, highlighting the critical need for new defensive mechanisms against targeted attacks within FL for autonomous driving.
Authors: Hiroki Waida, Kimihiro Yamazaki, Atsushi Tokuhisa, Mutsuyo Wada, Yuichiro Wada
Abstract: Self-supervised learning for image denoising problems in the presence of denaturation for noisy data is a crucial approach in machine learning. However, theoretical understanding of the performance of the approach that uses denatured data is lacking. To provide better understanding of the approach, in this paper, we analyze a self-supervised denoising algorithm that uses denatured data in depth through theoretical analysis and numerical experiments. Through the theoretical analysis, we discuss that the algorithm finds desired solutions to the optimization problem with the population risk, while the guarantee for the empirical risk depends on the hardness of the denoising task in terms of denaturation levels. We also conduct several experiments to investigate the performance of an extended algorithm in practice. The results indicate that the algorithm training with denatured images works, and the empirical performance aligns with the theoretical results. These results suggest several insights for further improvement of self-supervised image denoising that uses denatured data in future directions.
Authors: Abdallah Ayad, Adrian R\"ofer, Nick Heppert, Abhinav Valada
Abstract: Humans seemingly incorporate potential touch signals in their perception. Our goal is to equip robots with a similar capability, which we term Imagine2touch. Imagine2touch aims to predict the expected touch signal based on a visual patch representing the area to be touched. We use ReSkin, an inexpensive and compact touch sensor to collect the required dataset through random touching of five basic geometric shapes, and one tool. We train Imagine2touch on two out of those shapes and validate it on the ood. tool. We demonstrate the efficacy of Imagine2touch through its application to the downstream task of object recognition. In this task, we evaluate Imagine2touch performance in two experiments, together comprising 5 out of training distribution objects. Imagine2touch achieves an object recognition accuracy of 58% after ten touches per object, surpassing a proprioception baseline.
Authors: Pedro Mendes, Paolo Romano, David Garlan
Abstract: Neural networks are often overconfident about their predictions, which undermines their reliability and trustworthiness. In this work, we present a novel technique, named Error-Driven Uncertainty Aware Training (EUAT), which aims to enhance the ability of neural models to estimate their uncertainty correctly, namely to be highly uncertain when they output inaccurate predictions and low uncertain when their output is accurate. The EUAT approach operates during the model's training phase by selectively employing two loss functions depending on whether the training examples are correctly or incorrectly predicted by the model. This allows for pursuing the twofold goal of i) minimizing model uncertainty for correctly predicted inputs and ii) maximizing uncertainty for mispredicted inputs, while preserving the model's misprediction rate. We evaluate EUAT using diverse neural models and datasets in the image recognition domains considering both non-adversarial and adversarial settings. The results show that EUAT outperforms existing approaches for uncertainty estimation (including other uncertainty-aware training techniques, calibration, ensembles, and DEUP) by providing uncertainty estimates that not only have higher quality when evaluated via statistical metrics (e.g., correlation with residuals) but also when employed to build binary classifiers that decide whether the model's output can be trusted or not and under distributional data shifts.
Authors: Guangming Wang, Lei Pan, Songyou Peng, Shaohui Liu, Chenfeng Xu, Yanzi Miao, Wei Zhan, Masayoshi Tomizuka, Marc Pollefeys, Hesheng Wang
Abstract: Meticulous 3D environment representations have been a longstanding goal in computer vision and robotics fields. The recent emergence of neural implicit representations has introduced radical innovation to this field as implicit representations enable numerous capabilities. Among these, the Neural Radiance Field (NeRF) has sparked a trend because of the huge representational advantages, such as simplified mathematical models, compact environment storage, and continuous scene representations. Apart from computer vision, NeRF has also shown tremendous potential in the field of robotics. Thus, we create this survey to provide a comprehensive understanding of NeRF in the field of robotics. By exploring the advantages and limitations of NeRF, as well as its current applications and future potential, we hope to shed light on this promising area of research. Our survey is divided into two main sections: \textit{The Application of NeRF in Robotics} and \textit{The Advance of NeRF in Robotics}, from the perspective of how NeRF enters the field of robotics. In the first section, we introduce and analyze some works that have been or could be used in the field of robotics from the perception and interaction perspectives. In the second section, we show some works related to improving NeRF's own properties, which are essential for deploying NeRF in the field of robotics. In the discussion section of the review, we summarize the existing challenges and provide some valuable future research directions for reference.
Authors: Yi Yu, Yufei Wang, Song Xia, Wenhan Yang, Shijian Lu, Yap-Peng Tan, Alex C. Kot
Abstract: Unlearnable examples (UEs) seek to maximize testing error by making subtle modifications to training examples that are correctly labeled. Defenses against these poisoning attacks can be categorized based on whether specific interventions are adopted during training. The first approach is training-time defense, such as adversarial training, which can mitigate poisoning effects but is computationally intensive. The other approach is pre-training purification, e.g., image short squeezing, which consists of several simple compressions but often encounters challenges in dealing with various UEs. Our work provides a novel disentanglement mechanism to build an efficient pre-training purification method. Firstly, we uncover rate-constrained variational autoencoders (VAEs), demonstrating a clear tendency to suppress the perturbations in UEs. We subsequently conduct a theoretical analysis for this phenomenon. Building upon these insights, we introduce a disentangle variational autoencoder (D-VAE), capable of disentangling the perturbations with learnable class-wise embeddings. Based on this network, a two-stage purification approach is naturally developed. The first stage focuses on roughly eliminating perturbations, while the second stage produces refined, poison-free results, ensuring effectiveness and robustness across various scenarios. Extensive experiments demonstrate the remarkable performance of our method across CIFAR-10, CIFAR-100, and a 100-class ImageNet-subset. Code is available at https://github.com/yuyi-sd/D-VAE.
Authors: Yifei Ming, Yixuan Li
Abstract: Pre-trained contrastive vision-language models have demonstrated remarkable performance across a wide range of tasks. However, they often struggle on fine-trained datasets with categories not adequately represented during pre-training, which makes adaptation necessary. Recent works have shown promising results by utilizing samples from web-scale databases for retrieval-augmented adaptation, especially in low-data regimes. Despite the empirical success, understanding how retrieval impacts the adaptation of vision-language models remains an open research question. In this work, we adopt a reflective perspective by presenting a systematic study to understand the roles of key components in retrieval-augmented adaptation. We unveil new insights on uni-modal and cross-modal retrieval and highlight the critical role of logit ensemble for effective adaptation. We further present theoretical underpinnings that directly support our empirical observations.
Authors: Arkadiy Saakyan, Shreyas Kulkarni, Tuhin Chakrabarty, Smaranda Muresan
Abstract: Large Vision-Language models (VLMs) have demonstrated strong reasoning capabilities in tasks requiring a fine-grained understanding of literal images and text, such as visual question-answering or visual entailment. However, there has been little exploration of these models' capabilities when presented with images and captions containing figurative phenomena such as metaphors or humor, the meaning of which is often implicit. To close this gap, we propose a new task and a high-quality dataset: Visual Figurative Language Understanding with Textual Explanations (V-FLUTE). We frame the visual figurative language understanding problem as an explainable visual entailment task, where the model has to predict whether the image (premise) entails a claim (hypothesis) and justify the predicted label with a textual explanation. Using a human-AI collaboration framework, we build a high-quality dataset, V-FLUTE, that contains 6,027
Authors: Abhijit Das, Debesh Jha, Vandan Gorade, Koushik Biswas, Hongyi Pan, Zheyuan Zhang, Daniela P. Ladner, Yury Velichko, Amir Borhani, Ulas Bagci
Abstract: Computer-aided segmentation methods can assist medical personnel in improving diagnostic outcomes. While recent advancements like UNet and its variants have shown promise, they face a critical challenge: balancing accuracy with computational efficiency. Shallow encoder architectures in UNets often struggle to capture crucial spatial features, leading in inaccurate and sparse segmentation. To address this limitation, we propose a novel \underline{P}rogressive \underline{A}ttention based \underline{M}obile \underline{UNet} (\underline{PAM-UNet}) architecture. The inverted residual (IR) blocks in PAM-UNet help maintain a lightweight framework, while layerwise \textit{Progressive Luong Attention} ($\mathcal{PLA}$) promotes precise segmentation by directing attention toward regions of interest during synthesis. Our approach prioritizes both accuracy and speed, achieving a commendable balance with a mean IoU of 74.65 and a dice score of 82.87, while requiring only 1.32 floating-point operations per second (FLOPS) on the Liver Tumor Segmentation Benchmark (LiTS) 2017 dataset. These results highlight the importance of developing efficient segmentation models to accelerate the adoption of AI in clinical practice.
Authors: Luciano Dyballa, Evan Gerritz, Steven W. Zucker
Abstract: Generalization to unseen data remains poorly understood for deep learning classification and foundation models. How can one assess the ability of networks to adapt to new or extended versions of their input space in the spirit of few-shot learning, out-of-distribution generalization, and domain adaptation? Which layers of a network are likely to generalize best? We provide a new method for evaluating the capacity of networks to represent a sampled domain, regardless of whether the network has been trained on all classes in the domain. Our approach is the following: after fine-tuning state-of-the-art pre-trained models for visual classification on a particular domain, we assess their performance on data from related but distinct variations in that domain. Generalization power is quantified as a function of the latent embeddings of unseen data from intermediate layers for both unsupervised and supervised settings. Working throughout all stages of the network, we find that (i) high classification accuracy does not imply high generalizability; and (ii) deeper layers in a model do not always generalize the best, which has implications for pruning. Since the trends observed across datasets are largely consistent, we conclude that our approach reveals (a function of) the intrinsic capacity of the different layers of a model to generalize.
Authors: Homanga Bharadhwaj, Roozbeh Mottaghi, Abhinav Gupta, Shubham Tulsiani
Abstract: We seek to learn a generalizable goal-conditioned policy that enables zero-shot robot manipulation: interacting with unseen objects in novel scenes without test-time adaptation. While typical approaches rely on a large amount of demonstration data for such generalization, we propose an approach that leverages web videos to predict plausible interaction plans and learns a task-agnostic transformation to obtain robot actions in the real world. Our framework,Track2Act predicts tracks of how points in an image should move in future time-steps based on a goal, and can be trained with diverse videos on the web including those of humans and robots manipulating everyday objects. We use these 2D track predictions to infer a sequence of rigid transforms of the object to be manipulated, and obtain robot end-effector poses that can be executed in an open-loop manner. We then refine this open-loop plan by predicting residual actions through a closed loop policy trained with a few embodiment-specific demonstrations. We show that this approach of combining scalably learned track prediction with a residual policy requiring minimal in-domain robot-specific data enables zero-shot robot manipulation, and present a wide array of real-world robot manipulation results across unseen tasks, objects, and scenes. https://homangab.github.io/track2act/
Authors: Nishad Singhi, Jae Myung Kim, Karsten Roth, Zeynep Akata
Abstract: Concept Bottleneck Models (CBMs) ground image classification on human-understandable concepts to allow for interpretable model decisions. Crucially, the CBM design inherently allows for human interventions, in which expert users are given the ability to modify potentially misaligned concept choices to influence the decision behavior of the model in an interpretable fashion. However, existing approaches often require numerous human interventions per image to achieve strong performances, posing practical challenges in scenarios where obtaining human feedback is expensive. In this paper, we find that this is noticeably driven by an independent treatment of concepts during intervention, wherein a change of one concept does not influence the use of other ones in the model's final decision. To address this issue, we introduce a trainable concept intervention realignment module, which leverages concept relations to realign concept assignments post-intervention. Across standard, real-world benchmarks, we find that concept realignment can significantly improve intervention efficacy; significantly reducing the number of interventions needed to reach a target classification performance or concept prediction accuracy. In addition, it easily integrates into existing concept-based architectures without requiring changes to the models themselves. This reduced cost of human-model collaboration is crucial to enhancing the feasibility of CBMs in resource-constrained environments.
Authors: Murtaza Dalal, Tarun Chiruvolu, Devendra Chaplot, Ruslan Salakhutdinov
Abstract: Large Language Models (LLMs) have been shown to be capable of performing high-level planning for long-horizon robotics tasks, yet existing methods require access to a pre-defined skill library (e.g. picking, placing, pulling, pushing, navigating). However, LLM planning does not address how to design or learn those behaviors, which remains challenging particularly in long-horizon settings. Furthermore, for many tasks of interest, the robot needs to be able to adjust its behavior in a fine-grained manner, requiring the agent to be capable of modifying low-level control actions. Can we instead use the internet-scale knowledge from LLMs for high-level policies, guiding reinforcement learning (RL) policies to efficiently solve robotic control tasks online without requiring a pre-determined set of skills? In this paper, we propose Plan-Seq-Learn (PSL): a modular approach that uses motion planning to bridge the gap between abstract language and learned low-level control for solving long-horizon robotics tasks from scratch. We demonstrate that PSL achieves state-of-the-art results on over 25 challenging robotics tasks with up to 10 stages. PSL solves long-horizon tasks from raw visual input spanning four benchmarks at success rates of over 85%, out-performing language-based, classical, and end-to-end approaches. Video results and code at https://mihdalal.github.io/planseqlearn/
Authors: Yahav Avigal, Samuel Paradis, Harry Zhang
Abstract: Recent consumer demand for home robots has accelerated performance of robotic grasping. However, a key component of the perception pipeline, the depth camera, is still expensive and inaccessible to most consumers. In addition, grasp planning has significantly improved recently, by leveraging large datasets and cloud robotics, and by limiting the state and action space to top-down grasps with 4 degrees of freedom (DoF). By leveraging multi-view geometry of the object using inexpensive equipment such as off-the-shelf RGB cameras and state-of-the-art algorithms such as Learn Stereo Machine (LSM\cite{kar2017learning}), the robot is able to generate more robust grasps from different angles with 6-DoF. In this paper, we present a modification of LSM to graspable objects, evaluate the grasps, and develop a 6-DoF grasp planner based on Grasp-Quality CNN (GQ-CNN\cite{mahler2017dex}) that exploits multiple camera views to plan a robust grasp, even in the absence of a possible top-down grasp.
Authors: Brian Hsuan-Cheng Liao, Chih-Hong Cheng, Hasan Esen, Alois Knoll
Abstract: We consider the safety-oriented performance of 3D object detectors in autonomous driving contexts. Specifically, despite impressive results shown by the mass literature, developers often find it hard to ensure the safe deployment of these learning-based perception models. Attributing the challenge to the lack of safety-oriented metrics, we hereby present uncompromising spatial constraints (USC), which characterize a simple yet important localization requirement demanding the predictions to fully cover the objects when seen from the autonomous vehicle. The constraints, as we formulate using the perspective and bird's-eye views, can be naturally reflected by quantitative measures, such that having an object detector with a higher score implies a lower risk of collision. Finally, beyond model evaluation, we incorporate the quantitative measures into common loss functions to enable safety-oriented fine-tuning for existing models. With experiments using the nuScenes dataset and a closed-loop simulation, our work demonstrates such considerations of safety notions at the perception level not only improve model performances beyond accuracy but also allow for a more direct linkage to actual system safety.
Authors: Xiaoqing Zhang, Zunjie Xiao, Xiao Wu, Yanlin Chen, Jilu Zhao, Yan Hu, Jiang Liu
Abstract: Spatial attention mechanism has been widely incorporated into deep neural networks (DNNs), significantly lifting the performance in computer vision tasks via long-range dependency modeling. However, it may perform poorly in medical image analysis. Unfortunately, existing efforts are often unaware that long-range dependency modeling has limitations in highlighting subtle lesion regions. To overcome this limitation, we propose a practical yet lightweight architectural unit, Pyramid Pixel Context Adaption (PPCA) module, which exploits multi-scale pixel context information to recalibrate pixel position in a pixel-independent manner dynamically. PPCA first applies a well-designed cross-channel pyramid pooling to aggregate multi-scale pixel context information, then eliminates the inconsistency among them by the well-designed pixel normalization, and finally estimates per pixel attention weight via a pixel context integration. By embedding PPCA into a DNN with negligible overhead, the PPCANet is developed for medical image classification. In addition, we introduce supervised contrastive learning to enhance feature representation by exploiting the potential of label information via supervised contrastive loss. The extensive experiments on six medical image datasets show that PPCANet outperforms state-of-the-art attention-based networks and recent deep neural networks. We also provide visual analysis and ablation study to explain the behavior of PPCANet in the decision-making process.
Authors: Gemma Canet Tarr\'es, Dan Ruta, Tu Bui, John Collomosse
Abstract: We propose PARASOL, a multi-modal synthesis model that enables disentangled, parametric control of the visual style of the image by jointly conditioning synthesis on both content and a fine-grained visual style embedding. We train a latent diffusion model (LDM) using specific losses for each modality and adapt the classifier-free guidance for encouraging disentangled control over independent content and style modalities at inference time. We leverage auxiliary semantic and style-based search to create training triplets for supervision of the LDM, ensuring complementarity of content and style cues. PARASOL shows promise for enabling nuanced control over visual style in diffusion models for image creation and stylization, as well as generative search where text-based search results may be adapted to more closely match user intent by interpolating both content and style descriptors.
Authors: James Seale Smith, Yen-Chang Hsu, Lingyu Zhang, Ting Hua, Zsolt Kira, Yilin Shen, Hongxia Jin
Abstract: Recent works demonstrate a remarkable ability to customize text-to-image diffusion models while only providing a few example images. What happens if you try to customize such models using multiple, fine-grained concepts in a sequential (i.e., continual) manner? In our work, we show that recent state-of-the-art customization of text-to-image models suffer from catastrophic forgetting when new concepts arrive sequentially. Specifically, when adding a new concept, the ability to generate high quality images of past, similar concepts degrade. To circumvent this forgetting, we propose a new method, C-LoRA, composed of a continually self-regularized low-rank adaptation in cross attention layers of the popular Stable Diffusion model. Furthermore, we use customization prompts which do not include the word of the customized object (i.e., "person" for a human face dataset) and are initialized as completely random embeddings. Importantly, our method induces only marginal additional parameter costs and requires no storage of user data for replay. We show that C-LoRA not only outperforms several baselines for our proposed setting of text-to-image continual customization, which we refer to as Continual Diffusion, but that we achieve a new state-of-the-art in the well-established rehearsal-free continual learning setting for image classification. The high achieving performance of C-LoRA in two separate domains positions it as a compelling solution for a wide range of applications, and we believe it has significant potential for practical impact. Project page: https://jamessealesmith.github.io/continual-diffusion/
URLs: https://jamessealesmith.github.io/continual-diffusion/
Authors: Yesheng Zhang, Xu Zhao, Dahong Qian
Abstract: Feature matching is a crucial technique in computer vision. A unified perspective for this task is to treat it as a searching problem, aiming at an efficient search strategy to narrow the search space to point matches between images. One of the key aspects of search strategy is the search space, which in current approaches is not carefully defined, resulting in limited matching accuracy. This paper, thus, pays attention to the search space and proposes to set the initial search space for point matching as the matched image areas containing prominent semantic, named semantic area matches. This search space favors point matching by salient features and alleviates the accuracy limitation in recent Transformer-based matching methods. To achieve this search space, we introduce a hierarchical feature matching framework: Area to Point Matching (A2PM), to first find semantic area matches between images and later perform point matching on area matches. We further propose Semantic and Geometry Area Matching (SGAM) method to realize this framework, which utilizes semantic prior and geometry consistency to establish accurate area matches between images. By integrating SGAM with off-the-shelf state-of-the-art matchers, our method, adopting the A2PM framework, achieves encouraging precision improvements in massive point matching and pose estimation experiments.
Authors: Wenzhao Zhao, Barbara D. Wichtmann, Steffen Albert, Angelika Maurer, Frank G. Z\"ollner, Ulrike Attenberger, J\"urgen Hesser
Abstract: Group-equivariant convolutional neural networks (G-CNN) heavily rely on parameter sharing to increase CNN's data efficiency and performance. However, the parameter-sharing strategy greatly increases the computational burden for each added parameter, which hampers its application to deep neural network models. In this paper, we address these problems by proposing a non-parameter-sharing approach for group equivariant neural networks. The proposed methods adaptively aggregate a diverse range of filters by a weighted sum of stochastically augmented decomposed filters. We give theoretical proof about how the continuous group convolution can be approximated by our methods. Our method applies to both continuous and discrete groups, where the augmentation is implemented using Monte Carlo sampling and bootstrap resampling, respectively. We demonstrate that our methods serve as an efficient extension of standard CNN. Experiments on group equivariance tests show how our methods can achieve superior performance to parameter-sharing group equivariant networks. Experiments on image classification and image denoising tasks show that in certain scenarios, with a suitable set of filter bases, our method helps improve the performance of standard CNNs and build efficient lightweight image denoising networks. The code will be available at https://github.com/ZhaoWenzhao/MCG_CNN.
Authors: Shuai Zhao, Ruijie Quan, Linchao Zhu, Yi Yang
Abstract: Pre-trained vision-language models~(VLMs) are the de-facto foundation models for various downstream tasks. However, scene text recognition methods still prefer backbones pre-trained on a single modality, namely, the visual modality, despite the potential of VLMs to serve as powerful scene text readers. For example, CLIP can robustly identify regular (horizontal) and irregular (rotated, curved, blurred, or occluded) text in images. With such merits, we transform CLIP into a scene text reader and introduce CLIP4STR, a simple yet effective STR method built upon image and text encoders of CLIP. It has two encoder-decoder branches: a visual branch and a cross-modal branch. The visual branch provides an initial prediction based on the visual feature, and the cross-modal branch refines this prediction by addressing the discrepancy between the visual feature and text semantics. To fully leverage the capabilities of both branches, we design a dual predict-and-refine decoding scheme for inference. We scale CLIP4STR in terms of the model size, pre-training data, and training data, achieving state-of-the-art performance on 11 STR benchmarks. Additionally, a comprehensive empirical study is provided to enhance the understanding of the adaptation of CLIP to STR. We believe our method establishes a simple yet strong baseline for future STR research with VLMs.
Authors: Gongye Liu, Haoze Sun, Jiayi Li, Fei Yin, Yujiu Yang
Abstract: Diffusion models have recently demonstrated an impressive ability to address inverse problems in an unsupervised manner. While existing methods primarily focus on modifying the posterior sampling process, the potential of the forward process remains largely unexplored. In this work, we propose Shortcut Sampling for Diffusion(SSD), a novel approach for solving inverse problems in a zero-shot manner. Instead of initiating from random noise, the core concept of SSD is to find a specific transitional state that bridges the measurement image y and the restored image x. By utilizing the shortcut path of "input - transitional state - output", SSD can achieve precise restoration with fewer steps. To derive the transitional state during the forward process, we introduce Distortion Adaptive Inversion. Moreover, we apply back projection as additional consistency constraints during the generation process. Experimentally, we demonstrate SSD's effectiveness on multiple representative IR tasks. Our method achieves competitive results with only 30 NFEs compared to state-of-the-art zero-shot methods(100 NFEs) and outperforms them with 100 NFEs in certain tasks. Code is available at https://github.com/GongyeLiu/SSD
Authors: Mete Ahishali, Mehmet Yamac, Serkan Kiranyaz, Moncef Gabbouj
Abstract: In this work, we propose a novel approach called Operational Support Estimator Networks (OSENs) for the support estimation task. Support Estimation (SE) is defined as finding the locations of non-zero elements in sparse signals. By its very nature, the mapping between the measurement and sparse signal is a non-linear operation. Traditional support estimators rely on computationally expensive iterative signal recovery techniques to achieve such non-linearity. Contrary to the convolutional layers, the proposed OSEN approach consists of operational layers that can learn such complex non-linearities without the need for deep networks. In this way, the performance of non-iterative support estimation is greatly improved. Moreover, the operational layers comprise so-called generative super neurons with non-local kernels. The kernel location for each neuron/feature map is optimized jointly for the SE task during training. We evaluate the OSENs in three different applications: i. support estimation from Compressive Sensing (CS) measurements, ii. representation-based classification, and iii. learning-aided CS reconstruction where the output of OSENs is used as prior knowledge to the CS algorithm for enhanced reconstruction. Experimental results show that the proposed approach achieves computational efficiency and outperforms competing methods, especially at low measurement rates by significant margins. The software implementation is shared at https://github.com/meteahishali/OSEN.
Authors: Hu Zhang, Yanchen Li, Luziwei Leng, Kaiwei Che, Qian Liu, Qinghai Guo, Jianxing Liao, Ran Cheng
Abstract: Event-based sensors, distinguished by their high temporal resolution of 1 $\mathrm{\mu}\text{s}$ and a dynamic range of 120 $\text{dB}$, stand out as ideal tools for deployment in fast-paced settings like vehicles and drones. Traditional object detection techniques that utilize Artificial Neural Networks (ANNs) face challenges due to the sparse and asynchronous nature of the events these sensors capture. In contrast, Spiking Neural Networks (SNNs) offer a promising alternative, providing a temporal representation that is inherently aligned with event-based data. This paper explores the unique membrane potential dynamics of SNNs and their ability to modulate sparse events. We introduce an innovative spike-triggered adaptive threshold mechanism designed for stable training. Building on these insights, we present a specialized spiking feature pyramid network (SpikeFPN) optimized for automotive event-based object detection. Comprehensive evaluations demonstrate that SpikeFPN surpasses both traditional SNNs and advanced ANNs enhanced with attention mechanisms. Evidently, SpikeFPN achieves a mean Average Precision (mAP) of 0.477 on the GEN1 Automotive Detection (GAD) benchmark dataset, marking significant increases over the selected SNN baselines. Moreover, the efficient design of SpikeFPN ensures robust performance while optimizing computational resources, attributed to its innate sparse computation capabilities.
Authors: Yupu Yao, Shangqi Deng, Zihan Cao, Harry Zhang, Liang-Jian Deng
Abstract: Diffusion models have exhibited promising progress in video generation. However, they often struggle to retain consistent details within local regions across frames. One underlying cause is that traditional diffusion models approximate Gaussian noise distribution by utilizing predictive noise, without fully accounting for the impact of inherent information within the input itself. Additionally, these models emphasize the distinction between predictions and references, neglecting information intrinsic to the videos. To address this limitation, inspired by the self-attention mechanism, we propose a novel text-to-video (T2V) generation network structure based on diffusion models, dubbed Additional Perturbation for Latent noise with Adversarial training (APLA). Our approach only necessitates a single video as input and builds upon pre-trained stable diffusion networks. Notably, we introduce an additional compact network, known as the Video Generation Transformer (VGT). This auxiliary component is designed to extract perturbations from the inherent information contained within the input, thereby refining inconsistent pixels during temporal predictions. We leverage a hybrid architecture of transformers and convolutions to compensate for temporal intricacies, enhancing consistency between different frames within the video. Experiments demonstrate a noticeable improvement in the consistency of the generated videos both qualitatively and quantitatively.
Authors: Takami Sato, Justin Yue, Nanze Chen, Ningfei Wang, Qi Alfred Chen
Abstract: Denoising probabilistic diffusion models have shown breakthrough performance to generate more photo-realistic images or human-level illustrations than the prior models such as GANs. This high image-generation capability has stimulated the creation of many downstream applications in various areas. However, we find that this technology is actually a double-edged sword: We identify a new type of attack, called the Natural Denoising Diffusion (NDD) attack based on the finding that state-of-the-art deep neural network (DNN) models still hold their prediction even if we intentionally remove their robust features, which are essential to the human visual system (HVS), through text prompts. The NDD attack shows a significantly high capability to generate low-cost, model-agnostic, and transferable adversarial attacks by exploiting the natural attack capability in diffusion models. To systematically evaluate the risk of the NDD attack, we perform a large-scale empirical study with our newly created dataset, the Natural Denoising Diffusion Attack (NDDA) dataset. We evaluate the natural attack capability by answering 6 research questions. Through a user study, we find that it can achieve an 88% detection rate while being stealthy to 93% of human subjects; we also find that the non-robust features embedded by diffusion models contribute to the natural attack capability. To confirm the model-agnostic and transferable attack capability, we perform the NDD attack against the Tesla Model 3 and find that 73% of the physically printed attacks can be detected as stop signs. Our hope is that the study and dataset can help our community be aware of the risks in diffusion models and facilitate further research toward robust DNN models.
Authors: Saeid Asgari Taghanaki, Aliasghar Khani, Ali Saheb Pasand, Amir Khasahmadi, Aditya Sanghi, Karl D. D. Willis, Ali Mahdavi-Amiri
Abstract: Interpreting the learned features of vision models has posed a longstanding challenge in the field of machine learning. To address this issue, we propose a novel method that leverages the capabilities of language models to interpret the learned features of pre-trained image classifiers. Our method, called TExplain, tackles this task by training a neural network to establish a connection between the feature space of image classifiers and language models. Then, during inference, our approach generates a vast number of sentences to explain the features learned by the classifier for a given image. These sentences are then used to extract the most frequent words, providing a comprehensive understanding of the learned features and patterns within the classifier. Our method, for the first time, utilizes these frequent words corresponding to a visual representation to provide insights into the decision-making process of the independently trained classifier, enabling the detection of spurious correlations, biases, and a deeper comprehension of its behavior. To validate the effectiveness of our approach, we conduct experiments on diverse datasets, including ImageNet-9L and Waterbirds. The results demonstrate the potential of our method to enhance the interpretability and robustness of image classifiers.
Authors: Minsik Jeon, Junwon Seo, Jihong Min
Abstract: Despite the success of deep learning-based object detection methods in recent years, it is still challenging to make the object detector reliable in adverse weather conditions such as rain and snow. For the robust performance of object detectors, unsupervised domain adaptation has been utilized to adapt the detection network trained on clear weather images to adverse weather images. While previous methods do not explicitly address weather corruption during adaptation, the domain gap between clear and adverse weather can be decomposed into two factors with distinct characteristics: a style gap and a weather gap. In this paper, we present an unsupervised domain adaptation framework for object detection that can more effectively adapt to real-world environments with adverse weather conditions by addressing these two gaps separately. Our method resolves the style gap by concentrating on style-related information of high-level features using an attention module. Using self-supervised contrastive learning, our framework then reduces the weather gap and acquires instance features that are robust to weather corruption. Extensive experiments demonstrate that our method outperforms other methods for object detection in adverse weather conditions.
Authors: Yuan-Ming Li, Ling-An Zeng, Jing-Ke Meng, Wei-Shi Zheng
Abstract: Action Quality Assessment (AQA) is a task that tries to answer how well an action is carried out. While remarkable progress has been achieved, existing works on AQA assume that all the training data are visible for training at one time, but do not enable continual learning on assessing new technical actions. In this work, we address such a Continual Learning problem in AQA (Continual-AQA), which urges a unified model to learn AQA tasks sequentially without forgetting. Our idea for modeling Continual-AQA is to sequentially learn a task-consistent score-discriminative feature distribution, in which the latent features express a strong correlation with the score labels regardless of the task or action types.From this perspective, we aim to mitigate the forgetting in Continual-AQA from two aspects. Firstly, to fuse the features of new and previous data into a score-discriminative distribution, a novel Feature-Score Correlation-Aware Rehearsal is proposed to store and reuse data from previous tasks with limited memory size. Secondly, an Action General-Specific Graph is developed to learn and decouple the action-general and action-specific knowledge so that the task-consistent score-discriminative features can be better extracted across various tasks. Extensive experiments are conducted to evaluate the contributions of proposed components. The comparisons with the existing continual learning methods additionally verify the effectiveness and versatility of our approach. Data and code are available at https://github.com/iSEE-Laboratory/Continual-AQA.
Authors: Robert J\"ochl, Andreas Uhl
Abstract: In the context of temporal image forensics, it is not evident that a neural network, trained on images from different time-slots (classes), exploits solely image age related features. Usually, images taken in close temporal proximity (e.g., belonging to the same age class) share some common content properties. Such content bias can be exploited by a neural network. In this work, a novel approach is proposed that evaluates the influence of image content. This approach is verified using synthetic images (where content bias can be ruled out) with an age signal embedded. Based on the proposed approach, it is shown that a deep learning approach proposed in the context of age classification is most likely highly dependent on the image content. As a possible countermeasure, two different models from the field of image steganalysis, along with three different preprocessing techniques to increase the signal-to-noise ratio (age signal to image content), are evaluated using the proposed method.
Authors: Jindong Gu, Xiaojun Jia, Pau de Jorge, Wenqain Yu, Xinwei Liu, Avery Ma, Yuan Xun, Anjun Hu, Ashkan Khakzar, Zhijiang Li, Xiaochun Cao, Philip Torr
Abstract: The emergence of Deep Neural Networks (DNNs) has revolutionized various domains by enabling the resolution of complex tasks spanning image recognition, natural language processing, and scientific problem-solving. However, this progress has also brought to light a concerning vulnerability: adversarial examples. These crafted inputs, imperceptible to humans, can manipulate machine learning models into making erroneous predictions, raising concerns for safety-critical applications. An intriguing property of this phenomenon is the transferability of adversarial examples, where perturbations crafted for one model can deceive another, often with a different architecture. This intriguing property enables black-box attacks which circumvents the need for detailed knowledge of the target model. This survey explores the landscape of the adversarial transferability of adversarial examples. We categorize existing methodologies to enhance adversarial transferability and discuss the fundamental principles guiding each approach. While the predominant body of research primarily concentrates on image classification, we also extend our discussion to encompass other vision tasks and beyond. Challenges and opportunities are discussed, highlighting the importance of fortifying DNNs against adversarial vulnerabilities in an evolving landscape.
Authors: Tony Lindeberg
Abstract: The influence of natural image transformations on receptive field responses is crucial for modelling visual operations in computer vision and biological vision. In this regard, covariance properties with respect to geometric image transformations in the earliest layers of the visual hierarchy are essential for expressing robust image operations, and for formulating invariant visual operations at higher levels. This paper defines and proves a set of joint covariance properties under compositions of spatial scaling transformations, spatial affine transformations, Galilean transformations and temporal scaling transformations, which make it possible to characterize how different types of image transformations interact with each other and the associated spatio-temporal receptive field responses. In this regard, we also extend the notion of scale-normalized derivatives to affine-normalized derivatives, to be able to obtain true affine-covariant properties of spatial derivatives, that are computed based on spatial smoothing with affine Gaussian kernels. The derived relations show how the parameters of the receptive fields need to be transformed, in order to match the output from spatio-temporal receptive fields under composed spatio-temporal image transformations. As a side effect, the presented proof for the joint covariance property over the integrated combination of the different geometric image transformations also provides specific proofs for the individual transformation properties, which have not previously been fully reported in the literature. The paper also presents an in-depth theoretical analysis of geometric interpretations of the derived covariance properties, as well as outlines a number of biological interpretations of these results.
Authors: Yongjie Duan, Zhiyu Pan, Jianjiang Feng, Jie Zhou
Abstract: Compared to minutia-based fingerprint representations, fixed-length representations are attractive due to simple and efficient matching. However, fixed-length fingerprint representations are limited in accuracy when matching fingerprints with different visible areas, which can occur due to different finger poses or acquisition methods. To address this issue, we propose a localized deep representation of fingerprint, named LDRF. By focusing on the discriminative characteristics within local regions, LDRF provides a more robust and accurate fixed-length representation for fingerprints with variable visible areas. LDRF can be adapted to retain information within any valid area, making it highly flexible. The matching scores produced by LDRF also exhibit intuitive statistical characteristics, which led us to propose a matching score normalization technique to mitigate the uncertainty in the cases of very small overlapping area. With this new technique, we can maintain a high level of accuracy and reliability in our fingerprint matching, even as the size of the database grows rapidly. Our experimental results on 21 datasets containing over 140K fingerprints of various finger poses and impression types show that LDRF outperforms other fixed-length representations and is robust to sensing technologies and impression types. Besides, the proposed matching score normalization effectively reduces the false match rate (FMR) in large-scale identification experiments comprising over 5.11 million fingerprints. Specifically, this technique results in a reduction of two orders of magnitude compared to matching without matching score normalization and five orders of magnitude compared to prior works.
Authors: Wieland Morgenstern, Florian Barthel, Anna Hilsmann, Peter Eisert
Abstract: 3D Gaussian Splatting has recently emerged as a highly promising technique for modeling of static 3D scenes. In contrast to Neural Radiance Fields, it utilizes efficient rasterization allowing for very fast rendering at high-quality. However, the storage size is significantly higher, which hinders practical deployment, e.g. on resource constrained devices. In this paper, we introduce a compact scene representation organizing the parameters of 3D Gaussian Splatting (3DGS) into a 2D grid with local homogeneity, ensuring a drastic reduction in storage requirements without compromising visual quality during rendering. Central to our idea is the explicit exploitation of perceptual redundancies present in natural scenes. In essence, the inherent nature of a scene allows for numerous permutations of Gaussian parameters to equivalently represent it. To this end, we propose a novel highly parallel algorithm that regularly arranges the high-dimensional Gaussian parameters into a 2D grid while preserving their neighborhood structure. During training, we further enforce local smoothness between the sorted parameters in the grid. The uncompressed Gaussians use the same structure as 3DGS, ensuring a seamless integration with established renderers. Our method achieves a reduction factor of 17x to 42x in size for complex scenes with no increase in training time, marking a substantial leap forward in the domain of 3D scene distribution and consumption. Additional information can be found on our project page: https://fraunhoferhhi.github.io/Self-Organizing-Gaussians/
URLs: https://fraunhoferhhi.github.io/Self-Organizing-Gaussians/
Authors: Taekyung Kim, Byeongho Heo, Dongyoon Han
Abstract: Masked image modeling (MIM) is a promising option for training Vision Transformers among various self-supervised learning (SSL) methods. The essence of MIM lies in token-wise masked token predictions, with targets tokenized from images or generated by pre-trained models such as vision-language models. While tokenizers or pre-trained models are plausible MIM targets, they often offer spatially inconsistent targets even for neighboring tokens, complicating models to learn unified discriminative representations. Our pilot study confirms that addressing spatial inconsistencies has the potential to enhance representation quality. Motivated by the findings, we introduce a novel self-supervision signal called Dynamic Token Morphing (DTM), which dynamically aggregates contextually related tokens to yield contextualized targets. DTM is compatible with various SSL frameworks; we showcase an improved MIM by employing DTM, barely introducing extra training costs. Our experiments on ImageNet-1K and ADE20K demonstrate the superiority of our methods compared with state-of-the-art, complex MIM methods. Furthermore, the comparative evaluation of the iNaturalists and fine-grained visual classification datasets further validates the transferability of our method on various downstream tasks. Code is available at https://github.com/naver-ai/dtm
Authors: Pascal Schlachter, Bin Yang
Abstract: In real-world applications, there is often a domain shift from training to test data. This observation resulted in the development of test-time adaptation (TTA). It aims to adapt a pre-trained source model to the test data without requiring access to the source data. Thereby, most existing works are limited to the closed-set assumption, i.e. there is no category shift between source and target domain. We argue that in a realistic open-world setting a category shift can appear in addition to a domain shift. This means, individual source classes may not appear in the target domain anymore, samples of new classes may be part of the target domain or even both at the same time. Moreover, in many real-world scenarios the test data is not accessible all at once but arrives sequentially as a stream of batches demanding an immediate prediction. Hence, TTA must be applied in an online manner. To the best of our knowledge, the combination of these aspects, i.e. online source-free universal domain adaptation (online SF-UniDA), has not been studied yet. In this paper, we introduce a Contrastive Mean Teacher (COMET) tailored to this novel scenario. It applies a contrastive loss to rebuild a feature space where the samples of known classes build distinct clusters and the samples of new classes separate well from them. It is complemented by an entropy loss which ensures that the classifier output has a small entropy for samples of known classes and a large entropy for samples of new classes to be easily detected and rejected as unknown. To provide the losses with reliable pseudo labels, they are embedded into a mean teacher (MT) framework. We evaluate our method across two datasets and all category shifts to set an initial benchmark for online SF-UniDA. Thereby, COMET yields state-of-the-art performance and proves to be consistent and robust across a variety of different scenarios.
Authors: Junyoung Seo, Susung Hong, Wooseok Jang, In\`es Hyeonsu Kim, Minseop Kwak, Doyup Lee, Seungryong Kim
Abstract: Text-to-3D generation has achieved significant success by incorporating powerful 2D diffusion models, but insufficient 3D prior knowledge also leads to the inconsistency of 3D geometry. Recently, since large-scale multi-view datasets have been released, fine-tuning the diffusion model on the multi-view datasets becomes a mainstream to solve the 3D inconsistency problem. However, it has confronted with fundamental difficulties regarding the limited quality and diversity of 3D data, compared with 2D data. To sidestep these trade-offs, we explore a retrieval-augmented approach tailored for score distillation, dubbed ReDream. We postulate that both expressiveness of 2D diffusion models and geometric consistency of 3D assets can be fully leveraged by employing the semantically relevant assets directly within the optimization process. To this end, we introduce novel framework for retrieval-based quality enhancement in text-to-3D generation. We leverage the retrieved asset to incorporate its geometric prior in the variational objective and adapt the diffusion model's 2D prior toward view consistency, achieving drastic improvements in both geometry and fidelity of generated scenes. We conduct extensive experiments to demonstrate that ReDream exhibits superior quality with increased geometric consistency. Project page is available at https://ku-cvlab.github.io/ReDream/.
Authors: Xingchang Huang, Corentin Sala\"un, Cristina Vasconcelos, Christian Theobalt, Cengiz \"Oztireli, Gurprit Singh
Abstract: Most of the existing diffusion models use Gaussian noise for training and sampling across all time steps, which may not optimally account for the frequency contents reconstructed by the denoising network. Despite the diverse applications of correlated noise in computer graphics, its potential for improving the training process has been underexplored. In this paper, we introduce a novel and general class of diffusion models taking correlated noise within and across images into account. More specifically, we propose a time-varying noise model to incorporate correlated noise into the training process, as well as a method for fast generation of correlated noise mask. Our model is built upon deterministic diffusion models and utilizes blue noise to help improve the generation quality compared to using Gaussian white (random) noise only. Further, our framework allows introducing correlation across images within a single mini-batch to improve gradient flow. We perform both qualitative and quantitative evaluations on a variety of datasets using our method, achieving improvements on different tasks over existing deterministic diffusion models in terms of FID metric.
Authors: Youngsik Yun, Jihie Kim
Abstract: Image Captioning generates descriptive sentences from images using Vision-Language Pre-trained models (VLPs) such as BLIP, which has improved greatly. However, current methods lack the generation of detailed descriptive captions for the cultural elements depicted in the images, such as the traditional clothing worn by people from Asian cultural groups. In this paper, we propose a new framework, \textbf{Culturally-aware Image Captioning (CIC)}, that generates captions and describes cultural elements extracted from cultural visual elements in images representing cultures. Inspired by methods combining visual modality and Large Language Models (LLMs) through appropriate prompts, our framework (1) generates questions based on cultural categories from images, (2) extracts cultural visual elements from Visual Question Answering (VQA) using generated questions, and (3) generates culturally-aware captions using LLMs with the prompts. Our human evaluation conducted on 45 participants from 4 different cultural groups with a high understanding of the corresponding culture shows that our proposed framework generates more culturally descriptive captions when compared to the image captioning baseline based on VLPs. Our code and dataset will be made publicly available upon acceptance.
Authors: Evan Gerritz, Luciano Dyballa, Steven W. Zucker
Abstract: Generalization to unseen data is a key desideratum for deep networks, but its relation to classification accuracy is unclear. Using a minimalist vision dataset and a measure of generalizability, we show that popular networks, from deep convolutional networks (CNNs) to transformers, vary in their power to extrapolate to unseen classes both across layers and across architectures. Accuracy is not a good predictor of generalizability, and generalization varies non-monotonically with layer depth.
Authors: Penghao Zhao, Hailin Zhang, Qinhan Yu, Zhengren Wang, Yunteng Geng, Fangcheng Fu, Ling Yang, Wentao Zhang, Jie Jiang, Bin Cui
Abstract: Advancements in model algorithms, the growth of foundational models, and access to high-quality datasets have propelled the evolution of Artificial Intelligence Generated Content (AIGC). Despite its notable successes, AIGC still faces hurdles such as updating knowledge, handling long-tail data, mitigating data leakage, and managing high training and inference costs. Retrieval-Augmented Generation (RAG) has recently emerged as a paradigm to address such challenges. In particular, RAG introduces the information retrieval process, which enhances the generation process by retrieving relevant objects from available data stores, leading to higher accuracy and better robustness. In this paper, we comprehensively review existing efforts that integrate RAG technique into AIGC scenarios. We first classify RAG foundations according to how the retriever augments the generator, distilling the fundamental abstractions of the augmentation methodologies for various retrievers and generators. This unified perspective encompasses all RAG scenarios, illuminating advancements and pivotal technologies that help with potential future progress. We also summarize additional enhancements methods for RAG, facilitating effective engineering and implementation of RAG systems. Then from another view, we survey on practical applications of RAG across different modalities and tasks, offering valuable references for researchers and practitioners. Furthermore, we introduce the benchmarks for RAG, discuss the limitations of current RAG systems, and suggest potential directions for future research. Github: https://github.com/PKU-DAIR/RAG-Survey.
Authors: Kira Wursthorn, Markus Hillemann, Markus Ulrich
Abstract: The estimation of 6D object poses is a fundamental task in many computer vision applications. Particularly, in high risk scenarios such as human-robot interaction, industrial inspection, and automation, reliable pose estimates are crucial. In the last years, increasingly accurate and robust deep-learning-based approaches for 6D object pose estimation have been proposed. Many top-performing methods are not end-to-end trainable but consist of multiple stages. In the context of deep uncertainty quantification, deep ensembles are considered as state of the art since they have been proven to produce well-calibrated and robust uncertainty estimates. However, deep ensembles can only be applied to methods that can be trained end-to-end. In this work, we propose a method to quantify the uncertainty of multi-stage 6D object pose estimation approaches with deep ensembles. For the implementation, we choose SurfEmb as representative, since it is one of the top-performing 6D object pose estimation approaches in the BOP Challenge 2022. We apply established metrics and concepts for deep uncertainty quantification to evaluate the results. Furthermore, we propose a novel uncertainty calibration score for regression tasks to quantify the quality of the estimated uncertainty.
Authors: Shuangjian Li, Tao Zhu, Furong Duan, Liming Chen, Huansheng Ning, Christopher Nugent, Yaping Wan
Abstract: Wearable sensor-based human activity recognition (HAR) is a critical research domain in activity perception. However, achieving high efficiency and long sequence recognition remains a challenge. Despite the extensive investigation of temporal deep learning models, such as CNNs, RNNs, and transformers, their extensive parameters often pose significant computational and memory constraints, rendering them less suitable for resource-constrained mobile health applications. This study introduces HARMamba, an innovative light-weight and versatile HAR architecture that combines selective bidirectional SSM and hardware-aware design. To optimize real-time resource consumption in practical scenarios, HARMamba employs linear recursive mechanisms and parameter discretization, allowing it to selectively focus on relevant input sequences while efficiently fusing scan and recompute operations. To address potential issues with invalid sensor data, the system processes the data stream through independent channels, dividing each channel into "patches" and appending classification token to the end of the sequence. Position embeddings are incorporated to represent the sequence order, and the activity categories are output through a classification header. The HARMamba Block serves as the fundamental component of the HARMamba architecture, enabling the effective capture of more discriminative activity sequence features. HARMamba outperforms contemporary state-of-the-art frameworks, delivering comparable or better accuracy with significantly reducing computational and memory demands. It's effectiveness has been extensively validated on public datasets like PAMAP2, WISDM, UNIMIB SHAR and UCI, showcasing impressive results.
Authors: Yixin Wan, Arjun Subramonian, Anaelia Ovalle, Zongyu Lin, Ashima Suvarna, Christina Chance, Hritik Bansal, Rebecca Pattichis, Kai-Wei Chang
Abstract: The recent advancement of large and powerful models with Text-to-Image (T2I) generation abilities -- such as OpenAI's DALLE-3 and Google's Gemini -- enables users to generate high-quality images from textual prompts. However, it has become increasingly evident that even simple prompts could cause T2I models to exhibit conspicuous social bias in generated images. Such bias might lead to both allocational and representational harms in society, further marginalizing minority groups. Noting this problem, a large body of recent works has been dedicated to investigating different dimensions of bias in T2I systems. However, an extensive review of these studies is lacking, hindering a systematic understanding of current progress and research gaps. We present the first extensive survey on bias in T2I generative models. In this survey, we review prior studies on dimensions of bias: Gender, Skintone, and Geo-Culture. Specifically, we discuss how these works define, evaluate, and mitigate different aspects of bias. We found that: (1) while gender and skintone biases are widely studied, geo-cultural bias remains under-explored; (2) most works on gender and skintone bias investigated occupational association, while other aspects are less frequently studied; (3) almost all gender bias works overlook non-binary identities in their studies; (4) evaluation datasets and metrics are scattered, with no unified framework for measuring biases; and (5) current mitigation methods fail to resolve biases comprehensively. Based on current limitations, we point out future research directions that contribute to human-centric definitions, evaluations, and mitigation of biases. We hope to highlight the importance of studying biases in T2I systems, as well as encourage future efforts to holistically understand and tackle biases, building fair and trustworthy T2I technologies for everyone.
Authors: Yeonguk Yu, Sungho Shin, Seunghyeok Back, Minhwan Ko, Sangjun Noh, Kyoobin Lee
Abstract: Test-time adaptation (TTA) aims to adapt a pre-trained model to a new test domain without access to source data after deployment. Existing approaches typically rely on self-training with pseudo-labels since ground-truth cannot be obtained from test data. Although the quality of pseudo labels is important for stable and accurate long-term adaptation, it has not been previously addressed. In this work, we propose DPLOT, a simple yet effective TTA framework that consists of two components: (1) domain-specific block selection and (2) pseudo-label generation using paired-view images. Specifically, we select blocks that involve domain-specific feature extraction and train these blocks by entropy minimization. After blocks are adjusted for current test domain, we generate pseudo-labels by averaging given test images and corresponding flipped counterparts. By simply using flip augmentation, we prevent a decrease in the quality of the pseudo-labels, which can be caused by the domain gap resulting from strong augmentation. Our experimental results demonstrate that DPLOT outperforms previous TTA methods in CIFAR10-C, CIFAR100-C, and ImageNet-C benchmarks, reducing error by up to 5.4%, 9.1%, and 2.9%, respectively. Also, we provide an extensive analysis to demonstrate effectiveness of our framework. Code is available at https://github.com/gist-ailab/domain-specific-block-selection-and-paired-view-pseudo-labeling-for-online-TTA.
Authors: Yiran Xu, Taesung Park, Richard Zhang, Yang Zhou, Eli Shechtman, Feng Liu, Jia-Bin Huang, Difan Liu
Abstract: Video super-resolution (VSR) approaches have shown impressive temporal consistency in upsampled videos. However, these approaches tend to generate blurrier results than their image counterparts as they are limited in their generative capability. This raises a fundamental question: can we extend the success of a generative image upsampler to the VSR task while preserving the temporal consistency? We introduce VideoGigaGAN, a new generative VSR model that can produce videos with high-frequency details and temporal consistency. VideoGigaGAN builds upon a large-scale image upsampler -- GigaGAN. Simply inflating GigaGAN to a video model by adding temporal modules produces severe temporal flickering. We identify several key issues and propose techniques that significantly improve the temporal consistency of upsampled videos. Our experiments show that, unlike previous VSR methods, VideoGigaGAN generates temporally consistent videos with more fine-grained appearance details. We validate the effectiveness of VideoGigaGAN by comparing it with state-of-the-art VSR models on public datasets and showcasing video results with $8\times$ super-resolution.
Authors: Haotian Xue, Yongxin Chen
Abstract: Adversarial examples for diffusion models are widely used as solutions for safety concerns. By adding adversarial perturbations to personal images, attackers can not edit or imitate them easily. However, it is essential to note that all these protections target the latent diffusion model (LDMs), the adversarial examples for diffusion models in the pixel space (PDMs) are largely overlooked. This may mislead us to think that the diffusion models are vulnerable to adversarial attacks like most deep models. In this paper, we show novel findings that: even though gradient-based white-box attacks can be used to attack the LDMs, they fail to attack PDMs. This finding is supported by extensive experiments of almost a wide range of attacking methods on various PDMs and LDMs with different model structures, which means diffusion models are indeed much more robust against adversarial attacks. We also find that PDMs can be used as an off-the-shelf purifier to effectively remove the adversarial patterns that were generated on LDMs to protect the images, which means that most protection methods nowadays, to some extent, cannot protect our images from malicious attacks. We hope that our insights will inspire the community to rethink the adversarial samples for diffusion models as protection methods and move forward to more effective protection. Codes are available in https://github.com/xavihart/PDM-Pure.
Authors: Ziyue Zhang, Mingbao Lin, Rongrong Ji
Abstract: We introduce ObjectAdd, a training-free diffusion modification method to add user-expected objects into user-specified area. The motive of ObjectAdd stems from: first, describing everything in one prompt can be difficult, and second, users often need to add objects into the generated image. To accommodate with real world, our ObjectAdd maintains accurate image consistency after adding objects with technical innovations in: (1) embedding-level concatenation to ensure correct text embedding coalesce; (2) object-driven layout control with latent and attention injection to ensure objects accessing user-specified area; (3) prompted image inpainting in an attention refocusing & object expansion fashion to ensure rest of the image stays the same. With a text-prompted image, our ObjectAdd allows users to specify a box and an object, and achieves: (1) adding object inside the box area; (2) exact content outside the box area; (3) flawless fusion between the two areas
Authors: Zhixiong Huang, Xinying Wang, Jinjiang Li, Shenglan Liu, Lin Feng
Abstract: Underwater scenes intrinsically involve degradation problems owing to heterogeneous ocean elements. Prevailing underwater image enhancement (UIE) methods stick to straightforward feature modeling to learn the mapping function, which leads to limited vision gain as it lacks more explicit physical cues (e.g., depth). In this work, we investigate injecting the depth prior into the deep UIE model for more precise scene enhancement capability. To this end, we present a novel depth-guided perception UIE framework, dubbed underwater variable zoom (UVZ). Specifically, UVZ resorts to a two-stage pipeline. First, a depth estimation network is designed to generate critical depth maps, combined with an auxiliary supervision network introduced to suppress estimation differences during training. Second, UVZ parses near-far scenarios by harnessing the predicted depth maps, enabling local and non-local perceiving in different regions. Extensive experiments on five benchmark datasets demonstrate that UVZ achieves superior visual gain and delivers promising quantitative metrics. Besides, UVZ is confirmed to exhibit good generalization in some visual tasks, especially in unusual lighting conditions. The code, models and results are available at: https://github.com/WindySprint/UVZ.
Authors: Tengjun Huang
Abstract: With the rise of Visual and Language Pretraining (VLP), an increasing number of downstream tasks are adopting the paradigm of pretraining followed by fine-tuning. Although this paradigm has demonstrated potential in various multimodal downstream tasks, its implementation in the remote sensing domain encounters some obstacles. Specifically, the tendency for same-modality embeddings to cluster together impedes efficient transfer learning. To tackle this issue, we review the aim of multimodal transfer learning for downstream tasks from a unified perspective, and rethink the optimization process based on three distinct objectives. We propose "Harmonized Transfer Learning and Modality Alignment (HarMA)", a method that simultaneously satisfies task constraints, modality alignment, and single-modality uniform alignment, while minimizing training overhead through parameter-efficient fine-tuning. Remarkably, without the need for external data for training, HarMA achieves state-of-the-art performance in two popular multimodal retrieval tasks in the field of remote sensing. Our experiments reveal that HarMA achieves competitive and even superior performance to fully fine-tuned models with only minimal adjustable parameters. Due to its simplicity, HarMA can be integrated into almost all existing multimodal pretraining models. We hope this method can facilitate the efficient application of large models to a wide range of downstream tasks while significantly reducing the resource consumption. Code is available at https://github.com/seekerhuang/HarMA.
Authors: Chenyang Liu, Keyan Chen, Bowen Chen, Haotian Zhang, Zhengxia Zou, Zhenwei Shi
Abstract: Remote Sensing Image Change Captioning (RSICC) aims to describe surface changes between multi-temporal remote sensing images in language, including the changed object categories, locations, and dynamics of changing objects (e.g., added or disappeared). This poses challenges to spatial and temporal modeling of bi-temporal features. Despite previous methods progressing in the spatial change perception, there are still weaknesses in joint spatial-temporal modeling. To address this, in this paper, we propose a novel RSCaMa model, which achieves efficient joint spatial-temporal modeling through multiple CaMa layers, enabling iterative refinement of bi-temporal features. To achieve efficient spatial modeling, we introduce the recently popular Mamba (a state space model) with a global receptive field and linear complexity into the RSICC task and propose the Spatial Difference-aware SSM (SD-SSM), overcoming limitations of previous CNN- and Transformer-based methods in the receptive field and computational complexity. SD-SSM enhances the model's ability to capture spatial changes sharply. In terms of efficient temporal modeling, considering the potential correlation between the temporal scanning characteristics of Mamba and the temporality of the RSICC, we propose the Temporal-Traversing SSM (TT-SSM), which scans bi-temporal features in a temporal cross-wise manner, enhancing the model's temporal understanding and information interaction. Experiments validate the effectiveness of the efficient joint spatial-temporal modeling and demonstrate the outstanding performance of RSCaMa and the potential of the Mamba in the RSICC task. Additionally, we systematically compare three different language decoders, including Mamba, GPT-style decoder, and Transformer decoder, providing valuable insights for future RSICC research. The code will be available at \emph{\url{https://github.com/Chen-Yang-Liu/RSCaMa}}
Authors: Anudeep Das, Vasisht Duddu, Rui Zhang, N. Asokan
Abstract: Diffusion-based text-to-image (T2I) models generate high-fidelity images for given textual prompts. They are trained on large datasets scraped from the Internet, potentially containing unacceptable concepts (e.g., copyright infringing or unsafe). Retraining T2I models after filtering out unacceptable concepts in the training data is inefficient and degrades utility. Hence, there is a need for concept removal techniques (CRTs) which are effective in removing unacceptable concepts, utility-preserving on acceptable concepts, and robust against evasion with adversarial prompts. None of the prior filtering and fine-tuning CRTs satisfy all these requirements simultaneously. We introduce Espresso, the first robust concept filter based on Contrastive Language-Image Pre-Training (CLIP). It identifies unacceptable concepts by projecting the generated image's embedding onto the vector connecting unacceptable and acceptable concepts in the joint text-image embedding space. This ensures robustness by restricting the adversary to adding noise only along this vector, in the direction of the acceptable concept. Further fine-tuning Espresso to separate embeddings of acceptable and unacceptable concepts, while preserving their pairing with image embeddings, ensures both effectiveness and utility. We evaluate Espresso on eleven concepts to show that it is effective (~5% CLIP accuracy on unacceptable concepts), utility-preserving (~93% normalized CLIP score on acceptable concepts), and robust (~4% CLIP accuracy on adversarial prompts for unacceptable concepts). Finally, we present theoretical bounds for the certified robustness of Espresso against adversarial prompts, and an empirical analysis.
Authors: Jie Hu, Yawen Huang, Yilin Lu, Guoyang Xie, Guannan Jiang, Yefeng Zheng, Zhichao Lu
Abstract: Anomaly synthesis is one of the effective methods to augment abnormal samples for training. However, current anomaly synthesis methods predominantly rely on texture information as input, which limits the fidelity of synthesized abnormal samples. Because texture information is insufficient to correctly depict the pattern of anomalies, especially for logical anomalies. To surmount this obstacle, we present the AnomalyXFusion framework, designed to harness multi-modality information to enhance the quality of synthesized abnormal samples. The AnomalyXFusion framework comprises two distinct yet synergistic modules: the Multi-modal In-Fusion (MIF) module and the Dynamic Dif-Fusion (DDF) module. The MIF module refines modality alignment by aggregating and integrating various modality features into a unified embedding space, termed X-embedding, which includes image, text, and mask features. Concurrently, the DDF module facilitates controlled generation through an adaptive adjustment of X-embedding conditioned on the diffusion steps. In addition, to reveal the multi-modality representational power of AnomalyXFusion, we propose a new dataset, called MVTec Caption. More precisely, MVTec Caption extends 2.2k accurate image-mask-text annotations for the MVTec AD and LOCO datasets. Comprehensive evaluations demonstrate the effectiveness of AnomalyXFusion, especially regarding the fidelity and diversity for logical anomalies. Project page: http:github.com/hujiecpp/MVTec-Caption
Authors: Chuer Pan, Brian Okorn, Harry Zhang, Ben Eisner, David Held
Abstract: How do we imbue robots with the ability to efficiently manipulate unseen objects and transfer relevant skills based on demonstrations? End-to-end learning methods often fail to generalize to novel objects or unseen configurations. Instead, we focus on the task-specific pose relationship between relevant parts of interacting objects. We conjecture that this relationship is a generalizable notion of a manipulation task that can transfer to new objects in the same category; examples include the relationship between the pose of a pan relative to an oven or the pose of a mug relative to a mug rack. We call this task-specific pose relationship "cross-pose" and provide a mathematical definition of this concept. We propose a vision-based system that learns to estimate the cross-pose between two objects for a given manipulation task using learned cross-object correspondences. The estimated cross-pose is then used to guide a downstream motion planner to manipulate the objects into the desired pose relationship (placing a pan into the oven or the mug onto the mug rack). We demonstrate our method's capability to generalize to unseen objects, in some cases after training on only 10 demonstrations in the real world. Results show that our system achieves state-of-the-art performance in both simulated and real-world experiments across a number of tasks. Supplementary information and videos can be found at https://sites.google.com/view/tax-pose/home.
Authors: Guy Ohayon, Tomer Michaeli, Michael Elad
Abstract: We study the behavior of deterministic methods for solving inverse problems in imaging. These methods are commonly designed to achieve two goals: (1) attaining high perceptual quality, and (2) generating reconstructions that are consistent with the measurements. We provide a rigorous proof that the better a predictor satisfies these two requirements, the larger its Lipschitz constant must be, regardless of the nature of the degradation involved. In particular, to approach perfect perceptual quality and perfect consistency, the Lipschitz constant of the model must grow to infinity. This implies that such methods are necessarily more susceptible to adversarial attacks. We demonstrate our theory on single image super-resolution algorithms, addressing both noisy and noiseless settings. We also show how this undesired behavior can be leveraged to explore the posterior distribution, thereby allowing the deterministic model to imitate stochastic methods.
Authors: Ziheng Zhao, Yao Zhang, Chaoyi Wu, Xiaoman Zhang, Ya Zhang, Yanfeng Wang, Weidi Xie
Abstract: In this study, we focus on building up a model that aims to Segment Anything in medical scenarios, driven by Text prompts, termed as SAT. Our main contributions are three folds: (i) for dataset construction, we combine multiple knowledge sources to construct the first multi-modal knowledge tree on human anatomy, including 6502 anatomical terminologies; Then we build up the largest and most comprehensive segmentation dataset for training, by collecting over 22K 3D medical image scans from 72 segmentation datasets with careful standardization on both image scans and label space; (ii) for architecture design, we formulate a universal segmentation model, that can be prompted by inputting medical terminologies in text form. We present knowledge-enhanced representation learning on the combination of a large number of datasets; (iii) for model evaluation, we train a SAT-Pro with only 447M parameters, to segment 72 different segmentation datasets with text prompt, resulting in 497 classes. We have thoroughly evaluated the model from three aspects: averaged by body regions, averaged by classes, and average by datasets, demonstrating comparable performance to 72 specialist nnU-Nets, i.e., we train nnU-Net models on each dataset/subset, resulting in 72 nnU-Nets with around 2.2B parameters for the 72 datasets. We will release all the codes, and models in this work.
Authors: Felix Ott, Lucas Heublein, Nisha Lakshmana Raichur, Tobias Feigl, Jonathan Hansen, Alexander R\"ugamer, Christopher Mutschler
Abstract: Jamming devices pose a significant threat by disrupting signals from the global navigation satellite system (GNSS), compromising the robustness of accurate positioning. Detecting anomalies in frequency snapshots is crucial to counteract these interferences effectively. The ability to adapt to diverse, unseen interference characteristics is essential for ensuring the reliability of GNSS in real-world applications. In this paper, we propose a few-shot learning (FSL) approach to adapt to new interference classes. Our method employs quadruplet selection for the model to learn representations using various positive and negative interference classes. Furthermore, our quadruplet variant selects pairs based on the aleatoric and epistemic uncertainty to differentiate between similar classes. We recorded a dataset at a motorway with eight interference classes on which our FSL method with quadruplet loss outperforms other FSL techniques in jammer classification accuracy with 97.66%. Dataset available at: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/FIOT_highway
URLs: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/FIOT_highway
Authors: Chaoqun Du, Yizeng Han, Gao Huang
Abstract: Recent advancements in semi-supervised learning have focused on a more realistic yet challenging task: addressing imbalances in labeled data while the class distribution of unlabeled data remains both unknown and potentially mismatched. Current approaches in this sphere often presuppose rigid assumptions regarding the class distribution of unlabeled data, thereby limiting the adaptability of models to only certain distribution ranges. In this study, we propose a novel approach, introducing a highly adaptable framework, designated as SimPro, which does not rely on any predefined assumptions about the distribution of unlabeled data. Our framework, grounded in a probabilistic model, innovatively refines the expectation-maximization (EM) algorithm by explicitly decoupling the modeling of conditional and marginal class distributions. This separation facilitates a closed-form solution for class distribution estimation during the maximization phase, leading to the formulation of a Bayes classifier. The Bayes classifier, in turn, enhances the quality of pseudo-labels in the expectation phase. Remarkably, the SimPro framework not only comes with theoretical guarantees but also is straightforward to implement. Moreover, we introduce two novel class distributions broadening the scope of the evaluation. Our method showcases consistent state-of-the-art performance across diverse benchmarks and data distribution scenarios. Our code is available at https://github.com/LeapLabTHU/SimPro.
Authors: Linzhi Wu, Xingyu Zhang, Yakun Zhang, Changyan Zheng, Tiejun Liu, Liang Xie, Ye Yan, Erwei Yin
Abstract: Lip reading, the process of interpreting silent speech from visual lip movements, has gained rising attention for its wide range of realistic applications. Deep learning approaches greatly improve current lip reading systems. However, lip reading in cross-speaker scenarios where the speaker identity changes, poses a challenging problem due to inter-speaker variability. A well-trained lip reading system may perform poorly when handling a brand new speaker. To learn a speaker-robust lip reading model, a key insight is to reduce visual variations across speakers, avoiding the model overfitting to specific speakers. In this work, in view of both input visual clues and latent representations based on a hybrid CTC/attention architecture, we propose to exploit the lip landmark-guided fine-grained visual clues instead of frequently-used mouth-cropped images as input features, diminishing speaker-specific appearance characteristics. Furthermore, a max-min mutual information regularization approach is proposed to capture speaker-insensitive latent representations. Experimental evaluations on public lip reading datasets demonstrate the effectiveness of the proposed approach under the intra-speaker and inter-speaker conditions.
Authors: Yiyang Ma, Wenhan Yang, Jiaying Liu
Abstract: The images produced by diffusion models can attain excellent perceptual quality. However, it is challenging for diffusion models to guarantee distortion, hence the integration of diffusion models and image compression models still needs more comprehensive explorations. This paper presents a diffusion-based image compression method that employs a privileged end-to-end decoder model as correction, which achieves better perceptual quality while guaranteeing the distortion to an extent. We build a diffusion model and design a novel paradigm that combines the diffusion model and an end-to-end decoder, and the latter is responsible for transmitting the privileged information extracted at the encoder side. Specifically, we theoretically analyze the reconstruction process of the diffusion models at the encoder side with the original images being visible. Based on the analysis, we introduce an end-to-end convolutional decoder to provide a better approximation of the score function $\nabla_{\mathbf{x}_t}\log p(\mathbf{x}_t)$ at the encoder side and effectively transmit the combination. Experiments demonstrate the superiority of our method in both distortion and perception compared with previous perceptual compression methods.
Authors: Tahmim Hossain, Sagor Chandro Bakchy
Abstract: A well-known retinal disease that sends blurry visions to the affected patients is Macular Degeneration. This research is based on classifying the healthy and macular degeneration fundus by localizing the affected region of the fundus. A CNN architecture and CNN with ResNet architecture (ResNet50, ResNet50v2, ResNet101, ResNet101v2, ResNet152, ResNet152v2) as the backbone are used to classify the two types of fundus. The data are split into three categories including (a) Training set is 90% and Testing set is 10% (b) Training set is 80% and Testing set is 20%, (c) Training set is 50% and Testing set is 50%. After the training, the best model has been selected from the evaluation metrics. Among the models, CNN with a backbone of ResNet50 performs best which gives the training accuracy of 98.7% for 90% train and 10% test data split. With this model, we have performed the Grad-CAM visualization to get the region of the affected area of the fundus.
Authors: Shengjie Ma, Yanlin Weng, Tianjia Shao, Kun Zhou
Abstract: We introduce 3D Gaussian blendshapes for modeling photorealistic head avatars. Taking a monocular video as input, we learn a base head model of neutral expression, along with a group of expression blendshapes, each of which corresponds to a basis expression in classical parametric face models. Both the neutral model and expression blendshapes are represented as 3D Gaussians, which contain a few properties to depict the avatar appearance. The avatar model of an arbitrary expression can be effectively generated by combining the neutral model and expression blendshapes through linear blending of Gaussians with the expression coefficients. High-fidelity head avatar animations can be synthesized in real time using Gaussian splatting. Compared to state-of-the-art methods, our Gaussian blendshape representation better captures high-frequency details exhibited in input video, and achieves superior rendering performance.
Authors: Trinath Sai Subhash Reddy Pittala, Uma Maheswara R Meleti, Manasa Thatipamula
Abstract: In this project, we have explored machine learning approaches for predicting hearing loss thresholds on the brain's gray matter 3D images. We have solved the problem statement in two phases. In the first phase, we used a 3D CNN model to reduce high-dimensional input into latent space and decode it into an original image to represent the input in rich feature space. In the second phase, we utilized this model to reduce input into rich features and used these features to train standard machine learning models for predicting hearing thresholds. We have experimented with autoencoders and variational autoencoders in the first phase for dimensionality reduction and explored random forest, XGBoost and multi-layer perceptron for regressing the thresholds. We split the given data set into training and testing sets and achieved an 8.80 range and 22.57 range for PT500 and PT4000 on the test set, respectively. We got the lowest RMSE using multi-layer perceptron among the other models. Our approach leverages the unique capabilities of VAEs to capture complex, non-linear relationships within high-dimensional neuroimaging data. We rigorously evaluated the models using various metrics, focusing on the root mean squared error (RMSE). The results highlight the efficacy of the multi-layer neural network model, which outperformed other techniques in terms of accuracy. This project advances the application of data mining in medical diagnostics and enhances our understanding of age-related hearing loss through innovative machine-learning frameworks.