Computer Vision and Pattern Recognition (cs.CV)

Abstract: Learning implicit templates as neural fields has recently shown impressive performance in unsupervised shape correspondence. Despite the success, we observe current approaches, which solely rely on geometric information, often learn suboptimal deformation across generic object shapes, which have high structural variability. In this paper, we highlight the importance of part deformation consistency and propose a semantic-aware implicit template learning framework to enable semantically plausible deformation. By leveraging semantic prior from a self-supervised feature extractor, we suggest local conditioning with novel semantic-aware deformation code and deformation consistency regularizations regarding part deformation, global deformation, and global scaling. Our extensive experiments demonstrate the superiority of the proposed method over baselines in various tasks: keypoint transfer, part label transfer, and texture transfer. More interestingly, our framework shows a larger performance gain under more challenging settings. We also provide qualitative analyses to validate the effectiveness of semantic-aware deformation. The code is available at https://github.com/mlvlab/PDC.

Abstract: We address a challenging lifelong few-shot image generation task for the first time. In this situation, a generative model learns a sequence of tasks using only a few samples per task. Consequently, the learned model encounters both catastrophic forgetting and overfitting problems at a time. Existing studies on lifelong GANs have proposed modulation-based methods to prevent catastrophic forgetting. However, they require considerable additional parameters and cannot generate high-fidelity and diverse images from limited data. On the other hand, the existing few-shot GANs suffer from severe catastrophic forgetting when learning multiple tasks. To alleviate these issues, we propose a framework called Lifelong Few-Shot GAN (LFS-GAN) that can generate high-quality and diverse images in lifelong few-shot image generation task. Our proposed framework learns each task using an efficient task-specific modulator - Learnable Factorized Tensor (LeFT). LeFT is rank-constrained and has a rich representation ability due to its unique reconstruction technique. Furthermore, we propose a novel mode seeking loss to improve the diversity of our model in low-data circumstances. Extensive experiments demonstrate that the proposed LFS-GAN can generate high-fidelity and diverse images without any forgetting and mode collapse in various domains, achieving state-of-the-art in lifelong few-shot image generation task. Surprisingly, we find that our LFS-GAN even outperforms the existing few-shot GANs in the few-shot image generation task. The code is available at Github.

Abstract: In this paper, we present a novel Amplitude-Modulated Stochastic Perturbation and Vortex Convolutional Network, AMSP-UOD, designed for underwater object detection. AMSP-UOD specifically addresses the impact of non-ideal imaging factors on detection accuracy in complex underwater environments. To mitigate the influence of noise on object detection performance, we propose AMSP Vortex Convolution (AMSP-VConv) to disrupt the noise distribution, enhance feature extraction capabilities, effectively reduce parameters, and improve network robustness. We design the Feature Association Decoupling Cross Stage Partial (FAD-CSP) module, which strengthens the association of long and short-range features, improving the network performance in complex underwater environments. Additionally, our sophisticated post-processing method, based on non-maximum suppression with aspect-ratio similarity thresholds, optimizes detection in dense scenes, such as waterweed and schools of fish, improving object detection accuracy. Extensive experiments on the URPC and RUOD datasets demonstrate that our method outperforms existing state-of-the-art methods in terms of accuracy and noise immunity. AMSP-UOD proposes an innovative solution with the potential for real-world applications. Code will be made publicly available.

Abstract: Interpretability is a crucial factor in building reliable models for various medical applications. Concept Bottleneck Models (CBMs) enable interpretable image classification by utilizing human-understandable concepts as intermediate targets. Unlike conventional methods that require extensive human labor to construct the concept set, recent works leveraging Large Language Models (LLMs) for generating concepts made automatic concept generation possible. However, those methods do not consider whether a concept is visually relevant or not, which is an important factor in computing meaningful concept scores. Therefore, we propose a visual activation score that measures whether the concept contains visual cues or not, which can be easily computed with unlabeled image data. Computed visual activation scores are then used to filter out the less visible concepts, thus resulting in a final concept set with visually meaningful concepts. Our experimental results show that adopting the proposed visual activation score for concept filtering consistently boosts performance compared to the baseline. Moreover, qualitative analyses also validate that visually relevant concepts are successfully selected with the visual activation score.

Abstract: In this work, we seek to predict camera poses across scenes with a multi-task learning manner, where we view the localization of each scene as a new task. We propose OFVL-MS, a unified framework that dispenses with the traditional practice of training a model for each individual scene and relieves gradient conflict induced by optimizing multiple scenes collectively, enabling efficient storage yet precise visual localization for all scenes. Technically, in the forward pass of OFVL-MS, we design a layer-adaptive sharing policy with a learnable score for each layer to automatically determine whether the layer is shared or not. Such sharing policy empowers us to acquire task-shared parameters for a reduction of storage cost and task-specific parameters for learning scene-related features to alleviate gradient conflict. In the backward pass of OFVL-MS, we introduce a gradient normalization algorithm that homogenizes the gradient magnitude of the task-shared parameters so that all tasks converge at the same pace. Furthermore, a sparse penalty loss is applied on the learnable scores to facilitate parameter sharing for all tasks without performance degradation. We conduct comprehensive experiments on multiple benchmarks and our new released indoor dataset LIVL, showing that OFVL-MS families significantly outperform the state-of-the-arts with fewer parameters. We also verify that OFVL-MS can generalize to a new scene with much few parameters while gaining superior localization performance.

Abstract: Visual restoration of underwater scenes is crucial for visual tasks, and avoiding interference from underwater media has become a prominent concern. In this work, we present a synergistic multiscale detail refinement via intrinsic supervision (SMDR-IS) to recover underwater scene details. The low-degradation stage provides multiscale detail for original stage, which achieves synergistic multiscale detail refinement through feature propagation via the adaptive selective intrinsic supervised feature module (ASISF), which achieves synergistic multiscale detail refinement. ASISF is developed using intrinsic supervision to precisely control and guide feature transmission in the multi-degradation stages. ASISF improves the multiscale detail refinement while reducing interference from irrelevant scene information from the low-degradation stage. Additionally, within the multi-degradation encoder-decoder of SMDR-IS, we introduce a bifocal intrinsic-context attention module (BICA). This module is designed to effectively leverage multi-scale scene information found in images, using intrinsic supervision principles as its foundation. BICA facilitates the guidance of higher-resolution spaces by leveraging lower-resolution spaces, considering the significant dependency of underwater image restoration on spatial contextual relationships. During the training process, the network gains advantages from the integration of a multi-degradation loss function. This function serves as a constraint, enabling the network to effectively exploit information across various scales. When compared with state-of-the-art methods, SMDR-IS demonstrates its outstanding performance. Code will be made publicly available.

Abstract: Recognizing target objects using an event-based camera draws more and more attention in recent years. Existing works usually represent the event streams into point-cloud, voxel, image, etc, and learn the feature representations using various deep neural networks. Their final results may be limited by the following factors: monotonous modal expressions and the design of the network structure. To address the aforementioned challenges, this paper proposes a novel dual-stream framework for event representation, extraction, and fusion. This framework simultaneously models two common representations: event images and event voxels. By utilizing Transformer and Structured Graph Neural Network (GNN) architectures, spatial information and three-dimensional stereo information can be learned separately. Additionally, a bottleneck Transformer is introduced to facilitate the fusion of the dual-stream information. Extensive experiments demonstrate that our proposed framework achieves state-of-the-art performance on two widely used event-based classification datasets. The source code of this work is available at: \url{https://github.com/Event-AHU/EFV_event_classification}

Abstract: Diffusion probabilistic models (DPMs) have been shown to generate high-quality images without the need for delicate adversarial training. However, the current sampling process in DPMs is prone to violent shaking. In this paper, we present a novel reverse sampler for DPMs inspired by the widely-used Adam optimizer. Our proposed sampler can be readily applied to a pre-trained diffusion model, utilizing momentum mechanisms and adaptive updating to smooth the reverse sampling process and ensure stable generation, resulting in outputs of enhanced quality. By implicitly reusing update directions from early steps, our proposed sampler achieves a better balance between high-level semantics and low-level details. Additionally, this sampler is flexible and can be easily integrated into pre-trained DPMs regardless of the sampler used during training. Our experimental results on multiple benchmarks demonstrate that our proposed reverse sampler yields remarkable improvements over different baselines. We will make the source code available.

Abstract: Dancing with music is always an essential human art form to express emotion. Due to the high temporal-spacial complexity, long-term 3D realist dance generation synchronized with music is challenging. Existing methods suffer from the freezing problem when generating long-term dances due to error accumulation and training-inference discrepancy. To address this, we design a conditional diffusion model, LongDanceDiff, for this sequence-to-sequence long-term dance generation, addressing the challenges of temporal coherency and spatial constraint. LongDanceDiff contains a transformer-based diffusion model, where the input is a concatenation of music, past motions, and noised future motions. This partial noising strategy leverages the full-attention mechanism and learns the dependencies among music and past motions. To enhance the diversity of generated dance motions and mitigate the freezing problem, we introduce a mutual information minimization objective that regularizes the dependency between past and future motions. We also address common visual quality issues in dance generation, such as foot sliding and unsmooth motion, by incorporating spatial constraints through a Global-Trajectory Modulation (GTM) layer and motion perceptual losses, thereby improving the smoothness and naturalness of motion generation. Extensive experiments demonstrate a significant improvement in our approach over the existing state-of-the-art methods. We plan to release our codes and models soon.

Abstract: Diffusion Probabilistic Models (DPMs) have demonstrated substantial promise in image generation tasks but heavily rely on the availability of large amounts of training data. Previous works, like GANs, have tackled the limited data problem by transferring pre-trained models learned with sufficient data. However, those methods are hard to be utilized in DPMs since the distinct differences between DPM-based and GAN-based methods, showing in the unique iterative denoising process integral and the need for many timesteps with no-targeted noise in DPMs. In this paper, we propose a novel DPMs-based transfer learning method, TAN, to address the limited data problem. It includes two strategies: similarity-guided training, which boosts transfer with a classifier, and adversarial noise selection which adaptive chooses targeted noise based on the input image. Extensive experiments in the context of few-shot image generation tasks demonstrate that our method is not only efficient but also excels in terms of image quality and diversity when compared to existing GAN-based and DDPM-based methods.

Abstract: Image dehazing is quite challenging in dense-haze scenarios, where quite less original information remains in the hazy image. Though previous methods have made marvelous progress, they still suffer from information loss in content and color in dense-haze scenarios. The recently emerged Denoising Diffusion Probabilistic Model (DDPM) exhibits strong generation ability, showing potential for solving this problem. However, DDPM fails to consider the physics property of dehazing task, limiting its information completion capacity. In this work, we propose DehazeDDPM: A DDPM-based and physics-aware image dehazing framework that applies to complex hazy scenarios. Specifically, DehazeDDPM works in two stages. The former stage physically models the dehazing task with the Atmospheric Scattering Model (ASM), pulling the distribution closer to the clear data and endowing DehazeDDPM with fog-aware ability. The latter stage exploits the strong generation ability of DDPM to compensate for the haze-induced huge information loss, by working in conjunction with the physical modelling. Extensive experiments demonstrate that our method attains state-of-the-art performance on both synthetic and real-world hazy datasets.

Abstract: It is now possible to reconstruct dynamic human motion and shape from a sparse set of cameras using Neural Radiance Fields (NeRF) driven by an underlying skeleton. However, a challenge remains to model the deformation of cloth and skin in relation to skeleton pose. Unlike existing avatar models that are learned implicitly or rely on a proxy surface, our approach is motivated by the observation that different poses necessitate unique frequency assignments. Neglecting this distinction yields noisy artifacts in smooth areas or blurs fine-grained texture and shape details in sharp regions. We develop a two-branch neural network that is adaptive and explicit in the frequency domain. The first branch is a graph neural network that models correlations among body parts locally, taking skeleton pose as input. The second branch combines these correlation features to a set of global frequencies and then modulates the feature encoding. Our experiments demonstrate that our network outperforms state-of-the-art methods in terms of preserving details and generalization capabilities.

Abstract: Building scalable vision-language models to learn from diverse, multimodal data remains an open challenge. In this paper, we introduce an Efficient Vision-languagE foundation model, namely EVE, which is one unified multimodal Transformer pre-trained solely by one unified pre-training task. Specifically, EVE encodes both vision and language within a shared Transformer network integrated with modality-aware sparse Mixture-of-Experts (MoE) modules, which capture modality-specific information by selectively switching to different experts. To unify pre-training tasks of vision and language, EVE performs masked signal modeling on image-text pairs to reconstruct masked signals, i.e., image pixels and text tokens, given visible signals. This simple yet effective pre-training objective accelerates training by 3.5x compared to the model pre-trained with Image-Text Contrastive and Image-Text Matching losses. Owing to the combination of the unified architecture and pre-training task, EVE is easy to scale up, enabling better downstream performance with fewer resources and faster training speed. Despite its simplicity, EVE achieves state-of-the-art performance on various vision-language downstream tasks, including visual question answering, visual reasoning, and image-text retrieval.

Abstract: Text-driven localized editing of 3D objects is particularly difficult as locally mixing the original 3D object with the intended new object and style effects without distorting the object's form is not a straightforward process. To address this issue, we propose a novel NeRF-based model, Blending-NeRF, which consists of two NeRF networks: pretrained NeRF and editable NeRF. Additionally, we introduce new blending operations that allow Blending-NeRF to properly edit target regions which are localized by text. By using a pretrained vision-language aligned model, CLIP, we guide Blending-NeRF to add new objects with varying colors and densities, modify textures, and remove parts of the original object. Our extensive experiments demonstrate that Blending-NeRF produces naturally and locally edited 3D objects from various text prompts.

Abstract: Real-world point clouds usually suffer from incompleteness and display different poses. While current point cloud completion methods excel in reproducing complete point clouds with consistent poses as seen in the training set, their performance tends to be unsatisfactory when handling point clouds with diverse poses. We propose a network named Rotation-Invariant Completion Network (RICNet), which consists of two parts: a Dual Pipeline Completion Network (DPCNet) and an enhancing module. Firstly, DPCNet generates a coarse complete point cloud. The feature extraction module of DPCNet can extract consistent features, no matter if the input point cloud has undergone rotation or translation. Subsequently, the enhancing module refines the fine-grained details of the final generated point cloud. RICNet achieves better rotation invariance in feature extraction and incorporates structural relationships in man-made objects. To assess the performance of RICNet and existing methods on point clouds with various poses, we applied random transformations to the point clouds in the MVP dataset and conducted experiments on them. Our experiments demonstrate that RICNet exhibits superior completion performance compared to existing methods.

Abstract: Network calibration aims to accurately estimate the level of confidences, which is particularly important for employing deep neural networks in real-world systems. Recent approaches leverage mixup to calibrate the network's predictions during training. However, they do not consider the problem that mixtures of labels in mixup may not accurately represent the actual distribution of augmented samples. In this paper, we present RankMixup, a novel mixup-based framework alleviating the problem of the mixture of labels for network calibration. To this end, we propose to use an ordinal ranking relationship between raw and mixup-augmented samples as an alternative supervisory signal to the label mixtures for network calibration. We hypothesize that the network should estimate a higher level of confidence for the raw samples than the augmented ones (Fig.1). To implement this idea, we introduce a mixup-based ranking loss (MRL) that encourages lower confidences for augmented samples compared to raw ones, maintaining the ranking relationship. We also propose to leverage the ranking relationship among multiple mixup-augmented samples to further improve the calibration capability. Augmented samples with larger mixing coefficients are expected to have higher confidences and vice versa (Fig.1). That is, the order of confidences should be aligned with that of mixing coefficients. To this end, we introduce a novel loss, M-NDCG, in order to reduce the number of misaligned pairs of the coefficients and confidences. Extensive experimental results on standard benchmarks for network calibration demonstrate the effectiveness of RankMixup.

Abstract: Text-Pedestrian Image Retrieval aims to use the text describing pedestrian appearance to retrieve the corresponding pedestrian image. This task involves not only modality discrepancy, but also the challenge of the textual diversity of pedestrians with the same identity. At present, although existing research progress has been made in text-pedestrian image retrieval, these methods do not comprehensively consider the above-mentioned problems. Considering these, this paper proposes a progressive feature mining and external knowledge-assisted feature purification method. Specifically, we use a progressive mining mode to enable the model to mine discriminative features from neglected information, thereby avoiding the loss of discriminative information and improving the expression ability of features. In addition, to further reduce the negative impact of modal discrepancy and text diversity on cross-modal matching, we propose to use other sample knowledge of the same modality, i.e., external knowledge to enhance identity-consistent features and weaken identity-inconsistent features. This process purifies features and alleviates the interference caused by textual diversity and negative sample correlation features of the same modal. Extensive experiments on three challenging datasets demonstrate the effectiveness and superiority of the proposed method, and the retrieval performance even surpasses that of the large-scale model-based method on large-scale datasets.

Abstract: Image Quality Assessment (IQA) constitutes a fundamental task within the field of computer vision, yet it remains an unresolved challenge, owing to the intricate distortion conditions, diverse image contents, and limited availability of data. Recently, the community has witnessed the emergence of numerous large-scale pretrained foundation models, which greatly benefit from dramatically increased data and parameter capacities. However, it remains an open problem whether the scaling law in high-level tasks is also applicable to IQA task which is closely related to low-level clues. In this paper, we demonstrate that with proper injection of local distortion features, a larger pretrained and fixed foundation model performs better in IQA tasks. Specifically, for the lack of local distortion structure and inductive bias of vision transformer (ViT), alongside the large-scale pretrained ViT, we use another pretrained convolution neural network (CNN), which is well known for capturing the local structure, to extract multi-scale image features. Further, we propose a local distortion extractor to obtain local distortion features from the pretrained CNN and a local distortion injector to inject the local distortion features into ViT. By only training the extractor and injector, our method can benefit from the rich knowledge in the powerful foundation models and achieve state-of-the-art performance on popular IQA datasets, indicating that IQA is not only a low-level problem but also benefits from stronger high-level features drawn from large-scale pretrained models.

Abstract: RGB-D action and gesture recognition remain an interesting topic in human-centered scene understanding, primarily due to the multiple granularities and large variation in human motion. Although many RGB-D based action and gesture recognition approaches have demonstrated remarkable results by utilizing highly integrated spatio-temporal representations across multiple modalities (i.e., RGB and depth data), they still encounter several challenges. Firstly, vanilla 3D convolution makes it hard to capture fine-grained motion differences between local clips under different modalities. Secondly, the intricate nature of highly integrated spatio-temporal modeling can lead to optimization difficulties. Thirdly, duplicate and unnecessary information can add complexity and complicate entangled spatio-temporal modeling. To address the above issues, we propose an innovative heuristic architecture called Multi-stage Factorized Spatio-Temporal (MFST) for RGB-D action and gesture recognition. The proposed MFST model comprises a 3D Central Difference Convolution Stem (CDC-Stem) module and multiple factorized spatio-temporal stages. The CDC-Stem enriches fine-grained temporal perception, and the multiple hierarchical spatio-temporal stages construct dimension-independent higher-order semantic primitives. Specifically, the CDC-Stem module captures bottom-level spatio-temporal features and passes them successively to the following spatio-temporal factored stages to capture the hierarchical spatial and temporal features through the Multi- Scale Convolution and Transformer (MSC-Trans) hybrid block and Weight-shared Multi-Scale Transformer (WMS-Trans) block. The seamless integration of these innovative designs results in a robust spatio-temporal representation that outperforms state-of-the-art approaches on RGB-D action and gesture recognition datasets.

Abstract: Time of Flight (ToF) is a prevalent depth sensing technology in the fields of robotics, medical imaging, and non-destructive testing. Yet, ToF sensing faces challenges from complex ambient conditions making an inverse modelling from the sparse temporal information intractable. This paper highlights the potential of modern super-resolution techniques to learn varying surroundings for a reliable and accurate ToF detection. Unlike existing models, we tailor an architecture for sub-sample precise semi-global signal localization by combining super-resolution with an efficient residual contraction block to balance between fine signal details and large scale contextual information. We consolidate research on ToF by conducting a benchmark comparison against six state-of-the-art methods for which we employ two publicly available datasets. This includes the release of our SToF-Chirp dataset captured by an airborne ultrasound transducer. Results showcase the superior performance of our proposed StofNet in terms of precision, reliability and model complexity. Our code is available at https://github.com/hahnec/stofnet.

Abstract: Large-scale well-annotated datasets are of great importance for training an effective object detector. However, obtaining accurate bounding box annotations is laborious and demanding. Unfortunately, the resultant noisy bounding boxes could cause corrupt supervision signals and thus diminish detection performance. Motivated by the observation that the real ground-truth is usually situated in the aggregation region of the proposals assigned to a noisy ground-truth, we propose DIStribution-aware CalibratiOn (DISCO) to model the spatial distribution of proposals for calibrating supervision signals. In DISCO, spatial distribution modeling is performed to statistically extract the potential locations of objects. Based on the modeled distribution, three distribution-aware techniques, i.e., distribution-aware proposal augmentation (DA-Aug), distribution-aware box refinement (DA-Ref), and distribution-aware confidence estimation (DA-Est), are developed to improve classification, localization, and interpretability, respectively. Extensive experiments on large-scale noisy image datasets (i.e., Pascal VOC and MS-COCO) demonstrate that DISCO can achieve state-of-the-art detection performance, especially at high noise levels.

Abstract: Grounding textual expressions on scene objects from first-person views is a truly demanding capability in developing agents that are aware of their surroundings and behave following intuitive text instructions. Such capability is of necessity for glass-devices or autonomous robots to localize referred objects in the real-world. In the conventional referring expression comprehension tasks of images, however, datasets are mostly constructed based on the web-crawled data and don't reflect diverse real-world structures on the task of grounding textual expressions in diverse objects in the real world. Recently, a massive-scale egocentric video dataset of Ego4D was proposed. Ego4D covers around the world diverse real-world scenes including numerous indoor and outdoor situations such as shopping, cooking, walking, talking, manufacturing, etc. Based on egocentric videos of Ego4D, we constructed a broad coverage of the video-based referring expression comprehension dataset: RefEgo. Our dataset includes more than 12k video clips and 41 hours for video-based referring expression comprehension annotation. In experiments, we combine the state-of-the-art 2D referring expression comprehension models with the object tracking algorithm, achieving the video-wise referred object tracking even in difficult conditions: the referred object becomes out-of-frame in the middle of the video or multiple similar objects are presented in the video.

Abstract: The large-scale visual-language pre-trained model, Contrastive Language-Image Pre-training (CLIP), has significantly improved image captioning for scenarios without human-annotated image-caption pairs. Recent advanced CLIP-based image captioning without human annotations follows a text-only training paradigm, i.e., reconstructing text from shared embedding space. Nevertheless, these approaches are limited by the training/inference gap or huge storage requirements for text embeddings. Given that it is trivial to obtain images in the real world, we propose CLIP-guided text GAN (CgT-GAN), which incorporates images into the training process to enable the model to "see" real visual modality. Particularly, we use adversarial training to teach CgT-GAN to mimic the phrases of an external text corpus and CLIP-based reward to provide semantic guidance. The caption generator is jointly rewarded based on the caption naturalness to human language calculated from the GAN's discriminator and the semantic guidance reward computed by the CLIP-based reward module. In addition to the cosine similarity as the semantic guidance reward (i.e., CLIP-cos), we further introduce a novel semantic guidance reward called CLIP-agg, which aligns the generated caption with a weighted text embedding by attentively aggregating the entire corpus. Experimental results on three subtasks (ZS-IC, In-UIC and Cross-UIC) show that CgT-GAN outperforms state-of-the-art methods significantly across all metrics. Code is available at https://github.com/Lihr747/CgtGAN.

Abstract: Scene Graph Generation (SGG) as a critical task in image understanding, facing the challenge of head-biased prediction caused by the long-tail distribution of predicates. However, current unbiased SGG methods can easily prioritize improving the prediction of tail predicates while ignoring the substantial sacrifice in the prediction of head predicates, leading to a shift from head bias to tail bias. To address this issue, we propose a model-agnostic Head-Tail Collaborative Learning (HTCL) network that includes head-prefer and tail-prefer feature representation branches that collaborate to achieve accurate recognition of both head and tail predicates. We also propose a self-supervised learning approach to enhance the prediction ability of the tail-prefer feature representation branch by constraining tail-prefer predicate features. Specifically, self-supervised learning converges head predicate features to their class centers while dispersing tail predicate features as much as possible through contrast learning and head center loss. We demonstrate the effectiveness of our HTCL by applying it to various SGG models on VG150, Open Images V6 and GQA200 datasets. The results show that our method achieves higher mean Recall with a minimal sacrifice in Recall and achieves a new state-of-the-art overall performance. Our code is available at https://github.com/wanglei0618/HTCL.

Abstract: Fall detection is a vital task in health monitoring, as it allows the system to trigger an alert and therefore enabling faster interventions when a person experiences a fall. Although most previous approaches rely on standard RGB video data, such detailed appearance-aware monitoring poses significant privacy concerns. Depth sensors, on the other hand, are better at preserving privacy as they merely capture the distance of objects from the sensor or camera, omitting color and texture information. In this paper, we introduce a privacy-supporting solution that makes the RGB-trained model applicable in depth domain and utilizes depth data at test time for fall detection. To achieve cross-modal fall detection, we present an unsupervised RGB to Depth (RGB2Depth) cross-modal domain adaptation approach that leverages labelled RGB data and unlabelled depth data during training. Our proposed pipeline incorporates an intermediate domain module for feature bridging, modality adversarial loss for modality discrimination, classification loss for pseudo-labeled depth data and labeled source data, triplet loss that considers both source and target domains, and a novel adaptive loss weight adjustment method for improved coordination among various losses. Our approach achieves state-of-the-art results in the unsupervised RGB2Depth domain adaptation task for fall detection. Code is available at https://github.com/1015206533/privacy_supporting_fall_detection.

Abstract: The visual models pretrained on large-scale benchmarks encode general knowledge and prove effective in building more powerful representations for downstream tasks. Most existing approaches follow the fine-tuning paradigm, either by initializing or regularizing the downstream model based on the pretrained one. The former fails to retain the knowledge in the successive fine-tuning phase, thereby prone to be over-fitting, and the latter imposes strong constraints to the weights or feature maps of the downstream model without considering semantic drift, often incurring insufficient optimization. To deal with these issues, we propose a novel fine-tuning framework, namely distribution regularization with semantic calibration (DR-Tune). It employs distribution regularization by enforcing the downstream task head to decrease its classification error on the pretrained feature distribution, which prevents it from over-fitting while enabling sufficient training of downstream encoders. Furthermore, to alleviate the interference by semantic drift, we develop the semantic calibration (SC) module to align the global shape and class centers of the pretrained and downstream feature distributions. Extensive experiments on widely used image classification datasets show that DR-Tune consistently improves the performance when combing with various backbones under different pretraining strategies. Code is available at: https://github.com/weeknan/DR-Tune.

Abstract: Generative text-to-image models such as Stable Diffusion allow users to generate images based on a textual description, the prompt. Changing the prompt is still the primary means for the user to change a generated image as desired. However, changing the image by reformulating the prompt remains a difficult process of trial and error, which has led to the emergence of prompt engineering as a new field of research. We propose and analyze methods to change the embedding of a prompt directly instead of the prompt text. It allows for more fine-grained and targeted control that takes into account user intentions. Our approach treats the generative text-to-image model as a continuous function and passes gradients between the image space and the prompt embedding space. By addressing different user interaction problems, we can apply this idea in three scenarios: (1) Optimization of a metric defined in image space that could measure, for example, image style. (2) Assistance of users in creative tasks by enabling them to navigate the image space along a selection of directions of "near" prompt embeddings. (3) Changing the embedding of the prompt to include information that the user has seen in a particular seed but finds difficult to describe in the prompt. Our experiments demonstrate the feasibility of the described methods.

Abstract: Small farms contribute to a large share of the productive land in developing countries. In regions such as sub-Saharan Africa, where 80% of farms are small (under 2 ha in size), the task of mapping smallholder cropland is an important part of tracking sustainability measures such as crop productivity. However, the visually diverse and nuanced appearance of small farms has limited the effectiveness of traditional approaches to cropland mapping. Here we introduce a new approach based on the detection of harvest piles characteristic of many smallholder systems throughout the world. We present HarvestNet, a dataset for mapping the presence of farms in the Ethiopian regions of Tigray and Amhara during 2020-2023, collected using expert knowledge and satellite images, totaling 7k hand-labeled images and 2k ground collected labels. We also benchmark a set of baselines including SOTA models in remote sensing with our best models having around 80% classification performance on hand labelled data and 90%, 98% accuracy on ground truth data for Tigray, Amhara respectively. We also perform a visual comparison with a widely used pre-existing coverage map and show that our model detects an extra 56,621 hectares of cropland in Tigray. We conclude that remote sensing of harvest piles can contribute to more timely and accurate cropland assessments in food insecure region.

Abstract: Existing shadow detection datasets often contain missing or mislabeled shadows, which can hinder the performance of deep learning models trained directly on such data. To address this issue, we propose SILT, the Shadow-aware Iterative Label Tuning framework, which explicitly considers noise in shadow labels and trains the deep model in a self-training manner. Specifically, we incorporate strong data augmentations with shadow counterfeiting to help the network better recognize non-shadow regions and alleviate overfitting. We also devise a simple yet effective label tuning strategy with global-local fusion and shadow-aware filtering to encourage the network to make significant refinements on the noisy labels. We evaluate the performance of SILT by relabeling the test set of the SBU dataset and conducting various experiments. Our results show that even a simple U-Net trained with SILT can outperform all state-of-the-art methods by a large margin. When trained on SBU / UCF / ISTD, our network can successfully reduce the Balanced Error Rate by 25.2% / 36.9% / 21.3% over the best state-of-the-art method.

Abstract: RGB-Thermal (RGB-T) pedestrian detection aims to locate the pedestrians in RGB-T image pairs to exploit the complementation between the two modalities for improving detection robustness in extreme conditions. Most existing algorithms assume that the RGB-T image pairs are well registered, while in the real world they are not aligned ideally due to parallax or different field-of-view of the cameras. The pedestrians in misaligned image pairs may locate at different positions in two images, which results in two challenges: 1) how to achieve inter-modality complementation using spatially misaligned RGB-T pedestrian patches, and 2) how to recognize the unpaired pedestrians at the boundary. To deal with these issues, we propose a new paradigm for unregistered RGB-T pedestrian detection, which predicts two separate pedestrian locations in the RGB and thermal images, respectively. Specifically, we propose a cross-modality proposal-guided feature mining (CPFM) mechanism to extract the two precise fusion features for representing the pedestrian in the two modalities, even if the RGB-T image pair is unaligned. It enables us to effectively exploit the complementation between the two modalities. With the CPFM mechanism, we build a two-stream dense detector; it predicts the two pedestrian locations in the two modalities based on the corresponding fusion feature mined by the CPFM mechanism. Besides, we design a data augmentation method, named Homography, to simulate the discrepancy in scales and views between images. We also investigate two non-maximum suppression (NMS) methods for post-processing. Favorable experimental results demonstrate the effectiveness and robustness of our method in dealing with unregistered pedestrians with different shifts.

Abstract: Point cloud has a wide range of applications in areas such as autonomous driving, mapping, navigation, scene reconstruction, and medical imaging. Due to its great potentials in these applications, point cloud processing has gained great attention in the field of computer vision. Among various point cloud processing techniques, deep learning (DL) has become one of the mainstream and effective methods for tasks such as detection, segmentation and classification. To reduce overfitting during training DL models and improve model performance especially when the amount and/or diversity of training data are limited, augmentation is often crucial. Although various point cloud data augmentation methods have been widely used in different point cloud processing tasks, there are currently no published systematic surveys or reviews of these methods. Therefore, this article surveys and discusses these methods and categorizes them into a taxonomy framework. Through the comprehensive evaluation and comparison of the augmentation methods, this article identifies their potentials and limitations and suggests possible future research directions. This work helps researchers gain a holistic understanding of the current status of point cloud data augmentation and promotes its wider application and development.

Abstract: Group sparsity in Machine Learning (ML) encourages simpler, more interpretable models with fewer active parameter groups. This work aims to incorporate structured group sparsity into the shared parameters of a Multi-Task Learning (MTL) framework, to develop parsimonious models that can effectively address multiple tasks with fewer parameters while maintaining comparable or superior performance to a dense model. Sparsifying the model during training helps decrease the model's memory footprint, computation requirements, and prediction time during inference. We use channel-wise l1/l2 group sparsity in the shared layers of the Convolutional Neural Network (CNN). This approach not only facilitates the elimination of extraneous groups (channels) but also imposes a penalty on the weights, thereby enhancing the learning of all tasks. We compare the outcomes of single-task and multi-task experiments under group sparsity on two publicly available MTL datasets, NYU-v2 and CelebAMask-HQ. We also investigate how changing the sparsification degree impacts both the performance of the model and the sparsity of groups.

Abstract: Segmenting cells and tracking their motion over time is a common task in biomedical applications. However, predicting accurate instance-wise segmentation and cell motions from microscopy imagery remains a challenging task. Using microstructured environments for analyzing single cells in a constant flow of media adds additional complexity. While large-scale labeled microscopy datasets are available, we are not aware of any large-scale dataset, including both cells and microstructures. In this paper, we introduce the trapped yeast cell (TYC) dataset, a novel dataset for understanding instance-level semantics and motions of cells in microstructures. We release $105$ dense annotated high-resolution brightfield microscopy images, including about $19$k instance masks. We also release $261$ curated video clips composed of $1293$ high-resolution microscopy images to facilitate unsupervised understanding of cell motions and morphology. TYC offers ten times more instance annotations than the previously largest dataset, including cells and microstructures. Our effort also exceeds previous attempts in terms of microstructure variability, resolution, complexity, and capturing device (microscopy) variability. We facilitate a unified comparison on our novel dataset by introducing a standardized evaluation strategy. TYC and evaluation code are publicly available under CC BY 4.0 license.

Abstract: Models for fine-grained image classification tasks, where the difference between some classes can be extremely subtle and the number of samples per class tends to be low, are particularly prone to picking up background-related biases and demand robust methods to handle potential examples with out-of-distribution (OOD) backgrounds. To gain deeper insights into this critical problem, our research investigates the impact of background-induced bias on fine-grained image classification, evaluating standard backbone models such as Convolutional Neural Network (CNN) and Vision Transformers (ViT). We explore two masking strategies to mitigate background-induced bias: Early masking, which removes background information at the (input) image level, and late masking, which selectively masks high-level spatial features corresponding to the background. Extensive experiments assess the behavior of CNN and ViT models under different masking strategies, with a focus on their generalization to OOD backgrounds. The obtained findings demonstrate that both proposed strategies enhance OOD performance compared to the baseline models, with early masking consistently exhibiting the best OOD performance. Notably, a ViT variant employing GAP-Pooled Patch token-based classification combined with early masking achieves the highest OOD robustness.

Abstract: Facial landmark detection is an essential technology for driver status tracking and has been in demand for real-time estimations. As a landmark coordinate prediction, heatmap-based methods are known to achieve a high accuracy, and Lite-HRNet can achieve a fast estimation. However, with Lite-HRNet, the problem of a heavy computational cost of the fusion block, which connects feature maps with different resolutions, has yet to be solved. In addition, the strong output module used in HRNetV2 is not applied to Lite-HRNet. Given these problems, we propose a novel architecture called Lite-HRNet Plus. Lite-HRNet Plus achieves two improvements: a novel fusion block based on a channel attention and a novel output module with less computational intensity using multi-resolution feature maps. Through experiments conducted on two facial landmark datasets, we confirmed that Lite-HRNet Plus further improved the accuracy in comparison with conventional methods, and achieved a state-of-the-art accuracy with a computational complexity with the range of 10M FLOPs.

Abstract: Many practical systems for image-based surface reconstruction employ a stereo/multi-stereo paradigm, due to its ability to scale for large scenes and its ease of implementation for out-of-core operations. In this process, multiple and abundant depth maps from stereo matching must be combined and fused into a single, consistent, and clean point cloud. However, the noises and outliers caused by stereo matching and the heterogenous geometric errors of the poses present a challenge for existing fusion algorithms, since they mostly assume Gaussian errors and predict fused results based on data from local spatial neighborhoods, which may inherit uncertainties from multiple depths resulting in lowered accuracy. In this paper, we propose a novel depth fusion paradigm, that instead of numerically fusing points from multiple depth maps, selects the best depth map per point, and combines them into a single and clean point cloud. This paradigm, called select-and-combine (SAC), is achieved through modeling the point level fusion using local Markov Netlets, a micro-network over point across neighboring views for depth/view selection, followed by a Netlets collapse process for point combination. The Markov Netlets are optimized such that they can inherently leverage spatial consistencies among depth maps of neighboring views, thus they can address errors beyond Gaussian ones. Our experiment results show that our approach outperforms existing depth fusion approaches by increasing the F1 score that considers both accuracy and completeness by 2.07% compared to the best existing method. Finally, our approach generates clearer point clouds that are 18% less redundant while with a higher accuracy before fusion

Abstract: Conflating/stitching 2.5D raster digital surface models (DSM) into a large one has been a running practice in geoscience applications, however, conflating full-3D mesh models, such as those from oblique photogrammetry, is extremely challenging. In this letter, we propose a novel approach to address this challenge by conflating multiple full-3D oblique photogrammetric models into a single, and seamless mesh for high-resolution site modeling. Given two or more individually collected and created photogrammetric meshes, we first propose to create a virtual camera field (with a panoramic field of view) to incubate virtual spaces represented by Truncated Signed Distance Field (TSDF), an implicit volumetric field friendly for linear 3D fusion; then we adaptively leverage the truncated bound of meshes in TSDF to conflate them into a single and accurate full 3D site model. With drone-based 3D meshes, we show that our approach significantly improves upon traditional methods for model conflations, to drive new potentials to create excessively large and accurate full 3D mesh models in support of geoscience and environmental applications.

Abstract: This paper discusses the benefits of incorporating multimodal data for improving latent emotion recognition accuracy, focusing on micro-expression (ME) and physiological signals (PS). The proposed approach presents a novel multimodal learning framework that combines ME and PS, including a 1D separable and mixable depthwise inception network, a standardised normal distribution weighted feature fusion method, and depth/physiology guided attention modules for multimodal learning. Experimental results show that the proposed approach outperforms the benchmark method, with the weighted fusion method and guided attention modules both contributing to enhanced performance.

Abstract: Non-photorealistic videos are in demand with the wave of the metaverse, but lack of sufficient research studies. This work aims to take a step forward to understand how humans perceive non-photorealistic videos with eye fixation (\ie, saliency detection), which is critical for enhancing media production, artistic design, and game user experience. To fill in the gap of missing a suitable dataset for this research line, we present NPF-200, the first large-scale multi-modal dataset of purely non-photorealistic videos with eye fixations. Our dataset has three characteristics: 1) it contains soundtracks that are essential according to vision and psychological studies; 2) it includes diverse semantic content and videos are of high-quality; 3) it has rich motions across and within videos. We conduct a series of analyses to gain deeper insights into this task and compare several state-of-the-art methods to explore the gap between natural images and non-photorealistic data. Additionally, as the human attention system tends to extract visual and audio features with different frequencies, we propose a universal frequency-aware multi-modal non-photorealistic saliency detection model called NPSNet, demonstrating the state-of-the-art performance of our task. The results uncover strengths and weaknesses of multi-modal network design and multi-domain training, opening up promising directions for future works. {Our dataset and code can be found at \url{https://github.com/Yangziyu/NPF200}}.

Abstract: This paper presents IP-SLT, a simple yet effective framework for sign language translation (SLT). Our IP-SLT adopts a recurrent structure and enhances the semantic representation (prototype) of the input sign language video via an iterative refinement manner. Our idea mimics the behavior of human reading, where a sentence can be digested repeatedly, till reaching accurate understanding. Technically, IP-SLT consists of feature extraction, prototype initialization, and iterative prototype refinement. The initialization module generates the initial prototype based on the visual feature extracted by the feature extraction module. Then, the iterative refinement module leverages the cross-attention mechanism to polish the previous prototype by aggregating it with the original video feature. Through repeated refinement, the prototype finally converges to a more stable and accurate state, leading to a fluent and appropriate translation. In addition, to leverage the sequential dependence of prototypes, we further propose an iterative distillation loss to compress the knowledge of the final iteration into previous ones. As the autoregressive decoding process is executed only once in inference, our IP-SLT is ready to improve various SLT systems with acceptable overhead. Extensive experiments are conducted on public benchmarks to demonstrate the effectiveness of the IP-SLT.

Abstract: Analysis of player movements is a crucial subset of sports analysis. Existing player movement analysis methods use recorded videos after the match is over. In this work, we propose an end-to-end framework for player movement analysis for badminton matches on live broadcast match videos. We only use the visual inputs from the match and, unlike other approaches which use multi-modal sensor data, our approach uses only visual cues. We propose a method to calculate the on-court distance covered by both the players from the video feed of a live broadcast badminton match. To perform this analysis, we focus on the gameplay by removing replays and other redundant parts of the broadcast match. We then perform player tracking to identify and track the movements of both players in each frame. Finally, we calculate the distance covered by each player and the average speed with which they move on the court. We further show a heatmap of the areas covered by the player on the court which is useful for analyzing the gameplay of the player. Our proposed framework was successfully used to analyze live broadcast matches in real-time during the Premier Badminton League 2019 (PBL 2019), with commentators and broadcasters appreciating the utility.

Abstract: Out-of-distribution (OOD) detection refers to training the model on an in-distribution (ID) dataset to classify whether the input images come from unknown classes. Considerable effort has been invested in designing various OOD detection methods based on either convolutional neural networks or transformers. However, zero-shot OOD detection methods driven by CLIP, which only require class names for ID, have received less attention. This paper presents a novel method, namely CLIP saying "no" (\textbf{CLIPN}), which empowers the logic of saying "no" within CLIP. Our key motivation is to equip CLIP with the capability of distinguishing OOD and ID samples using positive-semantic prompts and negation-semantic prompts. Specifically, we design a novel learnable "no" prompt and a "no" text encoder to capture negation semantics within images. Subsequently, we introduce two loss functions: the image-text binary-opposite loss and the text semantic-opposite loss, which we use to teach CLIPN to associate images with "no" prompts, thereby enabling it to identify unknown samples. Furthermore, we propose two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from "no" prompts and the text encoder. Experimental results on 9 benchmark datasets (3 ID datasets and 6 OOD datasets) for the OOD detection task demonstrate that CLIPN, based on ViT-B-16, outperforms 7 well-used algorithms by at least 2.34\% and 11.64\% in terms of AUROC and FPR95 for zero-shot OOD detection on ImageNet-1K. Our CLIPN can serve as a solid foundation for effectively leveraging CLIP in downstream OOD tasks. The code is available on https://github.com/xmed-lab/CLIPN}{https://github.com/xmed-lab/CLIPN.

Abstract: Vision Transformer has demonstrated impressive success across various vision tasks. However, its heavy computation cost, which grows quadratically with respect to the token sequence length, largely limits its power in handling large feature maps. To alleviate the computation cost, previous works rely on either fine-grained self-attentions restricted to local small regions, or global self-attentions but to shorten the sequence length resulting in coarse granularity. In this paper, we propose a novel model, termed as Self-guided Transformer~(SG-Former), towards effective global self-attention with adaptive fine granularity. At the heart of our approach is to utilize a significance map, which is estimated through hybrid-scale self-attention and evolves itself during training, to reallocate tokens based on the significance of each region. Intuitively, we assign more tokens to the salient regions for achieving fine-grained attention, while allocating fewer tokens to the minor regions in exchange for efficiency and global receptive fields. The proposed SG-Former achieves performance superior to state of the art: our base size model achieves \textbf{84.7\%} Top-1 accuracy on ImageNet-1K, \textbf{51.2mAP} bbAP on CoCo, \textbf{52.7mIoU} on ADE20K surpassing the Swin Transformer by \textbf{+1.3\% / +2.7 mAP/ +3 mIoU}, with lower computation costs and fewer parameters. The code is available at \href{https://github.com/OliverRensu/SG-Former}{https://github.com/OliverRensu/SG-Former}

Abstract: Existing methods of multiple human parsing (MHP) apply statistical models to acquire underlying associations between images and labeled body parts. However, acquired associations often contain many spurious correlations that degrade model generalization, leading statistical models to be vulnerable to visually contextual variations in images (e.g., unseen image styles/external interventions). To tackle this, we present a causality inspired parsing paradigm termed CIParsing, which follows fundamental causal principles involving two causal properties for human parsing (i.e., the causal diversity and the causal invariance). Specifically, we assume that an input image is constructed by a mix of causal factors (the characteristics of body parts) and non-causal factors (external contexts), where only the former ones cause the generation process of human parsing.Since causal/non-causal factors are unobservable, a human parser in proposed CIParsing is required to construct latent representations of causal factors and learns to enforce representations to satisfy the causal properties. In this way, the human parser is able to rely on causal factors w.r.t relevant evidence rather than non-causal factors w.r.t spurious correlations, thus alleviating model degradation and yielding improved parsing ability. Notably, the CIParsing is designed in a plug-and-play fashion and can be integrated into any existing MHP models. Extensive experiments conducted on two widely used benchmarks demonstrate the effectiveness and generalizability of our method.

Abstract: The automatic analysis of chemical literature has immense potential to accelerate the discovery of new materials and drugs. Much of the critical information in patent documents and scientific articles is contained in figures, depicting the molecule structures. However, automatically parsing the exact chemical structure is a formidable challenge, due to the amount of detailed information, the diversity of drawing styles, and the need for training data. In this work, we introduce MolGrapher to recognize chemical structures visually. First, a deep keypoint detector detects the atoms. Second, we treat all candidate atoms and bonds as nodes and put them in a graph. This construct allows a natural graph representation of the molecule. Last, we classify atom and bond nodes in the graph with a Graph Neural Network. To address the lack of real training data, we propose a synthetic data generation pipeline producing diverse and realistic results. In addition, we introduce a large-scale benchmark of annotated real molecule images, USPTO-30K, to spur research on this critical topic. Extensive experiments on five datasets show that our approach significantly outperforms classical and learning-based methods in most settings. Code, models, and datasets are available.

Abstract: Since thermal imagery offers a unique modality to investigate pain, the U.S. National Institutes of Health (NIH) has collected a large and diverse set of cancer patient facial thermograms for AI-based pain research. However, differing angles from camera capture between thermal and visible sensors has led to misalignment between Visible-Thermal (VT) images. We modernize the classic computer vision task of image registration by applying and modifying a generative alignment algorithm to register VT cancer faces, without the need for a reference or alignment parameters. By registering VT faces, we demonstrate that the quality of thermal images produced in the generative AI downstream task of Visible-to-Thermal (V2T) image translation significantly improves up to 52.5\%, than without registration. Images in this paper have been approved by the NIH NCI for public dissemination.

Abstract: We present a method for teaching machines to understand and model the underlying spatial common sense of diverse human-object interactions in 3D in a self-supervised way. This is a challenging task, as there exist specific manifolds of the interactions that can be considered human-like and natural, but the human pose and the geometry of objects can vary even for similar interactions. Such diversity makes the annotating task of 3D interactions difficult and hard to scale, which limits the potential to reason about that in a supervised way. One way of learning the 3D spatial relationship between humans and objects during interaction is by showing multiple 2D images captured from different viewpoints when humans interact with the same type of objects. The core idea of our method is to leverage a generative model that produces high-quality 2D images from an arbitrary text prompt input as an "unbounded" data generator with effective controllability and view diversity. Despite its imperfection of the image quality over real images, we demonstrate that the synthesized images are sufficient to learn the 3D human-object spatial relations. We present multiple strategies to leverage the synthesized images, including (1) the first method to leverage a generative image model for 3D human-object spatial relation learning; (2) a framework to reason about the 3D spatial relations from inconsistent 2D cues in a self-supervised manner via 3D occupancy reasoning with pose canonicalization; (3) semantic clustering to disambiguate different types of interactions with the same object types; and (4) a novel metric to assess the quality of 3D spatial learning of interaction. Project Page: https://jellyheadandrew.github.io/projects/chorus