Computer Vision and Pattern Recognition (cs.CV)

Abstract: The SnakeCLEF2023 competition aims to the development of advanced algorithms for snake species identification through the analysis of images and accompanying metadata. This paper presents a method leveraging utilization of both images and metadata. Modern CNN models and strong data augmentation are utilized to learn better representation of images. To relieve the challenge of long-tailed distribution, seesaw loss is utilized in our method. We also design a light model to calculate prior probabilities using metadata features extracted from CLIP in post processing stage. Besides, we attach more importance to venomous species by assigning venomous species labels to some examples that model is uncertain about. Our method achieves 91.31% score of the final metric combined of F1 and other metrics on private leaderboard, which is the 1st place among the participators. The code is available at https://github.com/xiaoxsparraw/CLEF2023.

Abstract: Scene text image super-resolution (STISR), aiming to improve image quality while boosting downstream scene text recognition accuracy, has recently achieved great success. However, most existing methods treat the foreground (character regions) and background (non-character regions) equally in the forward process, and neglect the disturbance from the complex background, thus limiting the performance. To address these issues, in this paper, we propose a novel method LEMMA that explicitly models character regions to produce high-level text-specific guidance for super-resolution. To model the location of characters effectively, we propose the location enhancement module to extract character region features based on the attention map sequence. Besides, we propose the multi-modal alignment module to perform bidirectional visual-semantic alignment to generate high-quality prior guidance, which is then incorporated into the super-resolution branch in an adaptive manner using the proposed adaptive fusion module. Experiments on TextZoom and four scene text recognition benchmarks demonstrate the superiority of our method over other state-of-the-art methods. Code is available at https://github.com/csguoh/LEMMA.

Abstract: Semi-Supervised Semantic Segmentation (S4) aims to train a segmentation model with limited labeled images and a substantial volume of unlabeled images. To improve the robustness of representations, powerful methods introduce a pixel-wise contrastive learning approach in latent space (i.e., representation space) that aggregates the representations to their prototypes in a fully supervised manner. However, previous contrastive-based S4 methods merely rely on the supervision from the model's output (logits) in logit space during unlabeled training. In contrast, we utilize the outputs in both logit space and representation space to obtain supervision in a collaborative way. The supervision from two spaces plays two roles: 1) reduces the risk of over-fitting to incorrect semantic information in logits with the help of representations; 2) enhances the knowledge exchange between the two spaces. Furthermore, unlike previous approaches, we use the similarity between representations and prototypes as a new indicator to tilt training those under-performing representations and achieve a more efficient contrastive learning process. Results on two public benchmarks demonstrate the competitive performance of our method compared with state-of-the-art methods.

Abstract: Weakly supervised object localization (WSOL) remains challenging when learning object localization models from image category labels. Conventional methods that discriminatively train activation models ignore representative yet less discriminative object parts. In this study, we propose a generative prompt model (GenPromp), defining the first generative pipeline to localize less discriminative object parts by formulating WSOL as a conditional image denoising procedure. During training, GenPromp converts image category labels to learnable prompt embeddings which are fed to a generative model to conditionally recover the input image with noise and learn representative embeddings. During inference, enPromp combines the representative embeddings with discriminative embeddings (queried from an off-the-shelf vision-language model) for both representative and discriminative capacity. The combined embeddings are finally used to generate multi-scale high-quality attention maps, which facilitate localizing full object extent. Experiments on CUB-200-2011 and ILSVRC show that GenPromp respectively outperforms the best discriminative models by 5.2% and 5.6% (Top-1 Loc), setting a solid baseline for WSOL with the generative model. Code is available at https://github.com/callsys/GenPromp.

Abstract: Chest X-Ray (CXR) report generation is a promising approach to improving the efficiency of CXR interpretation. However, a significant increase in diagnostic accuracy is required before that can be realised. Motivated by this, we propose a framework that is more inline with a radiologist's workflow by considering longitudinal data. Here, the decoder is additionally conditioned on the report from the subject's previous imaging study via a prompt. We also propose a new reward for reinforcement learning based on CXR-BERT, which computes the similarity between reports. We conduct experiments on the MIMIC-CXR dataset. The results indicate that longitudinal data improves CXR report generation. CXR-BERT is also shown to be a promising alternative to the current state-of-the-art reward based on RadGraph. This investigation indicates that longitudinal CXR report generation can offer a substantial increase in diagnostic accuracy. Our Hugging Face model is available at: https://huggingface.co/aehrc/cxrmate and code is available at: https://github.com/aehrc/cxrmate.

Abstract: The vulnerability of deep neural networks to adversarial samples has been a major impediment to their broad applications, despite their success in various fields. Recently, some works suggested that adversarially-trained models emphasize the importance of low-frequency information to achieve higher robustness. While several attempts have been made to leverage this frequency characteristic, they have all faced the issue that applying low-pass filters directly to input images leads to irreversible loss of discriminative information and poor generalizability to datasets with distinct frequency features. This paper presents a plug-and-play module called the Frequency Preference Control Module that adaptively reconfigures the low- and high-frequency components of intermediate feature representations, providing better utilization of frequency in robust learning. Empirical studies show that our proposed module can be easily incorporated into any adversarial training framework, further improving model robustness across different architectures and datasets. Additionally, experiments were conducted to examine how the frequency bias of robust models impacts the adversarial training process and its final robustness, revealing interesting insights.

Abstract: Unsupervised domain adaptation (UDA) has increasingly gained interests for its capacity to transfer the knowledge learned from a labeled source domain to an unlabeled target domain. However, typical UDA methods require concurrent access to both the source and target domain data, which largely limits its application in medical scenarios where source data is often unavailable due to privacy concern. To tackle the source data-absent problem, we present a novel two-stage source-free domain adaptation (SFDA) framework for medical image segmentation, where only a well-trained source segmentation model and unlabeled target data are available during domain adaptation. Specifically, in the prototype-anchored feature alignment stage, we first utilize the weights of the pre-trained pixel-wise classifier as source prototypes, which preserve the information of source features. Then, we introduce the bi-directional transport to align the target features with class prototypes by minimizing its expected cost. On top of that, a contrastive learning stage is further devised to utilize those pixels with unreliable predictions for a more compact target feature distribution. Extensive experiments on a cross-modality medical segmentation task demonstrate the superiority of our method in large domain discrepancy settings compared with the state-of-the-art SFDA approaches and even some UDA methods. Code is available at https://github.com/CSCYQJ/MICCAI23-ProtoContra-SFDA.

Abstract: Layer compositing is one of the most popular image editing workflows among both amateurs and professionals. Motivated by the success of diffusion models, we explore layer compositing from a layered image generation perspective. Instead of generating an image, we propose to generate background, foreground, layer mask, and the composed image simultaneously. To achieve layered image generation, we train an autoencoder that is able to reconstruct layered images and train diffusion models on the latent representation. One benefit of the proposed problem is to enable better compositing workflows in addition to the high-quality image output. Another benefit is producing higher-quality layer masks compared to masks produced by a separate step of image segmentation. Experimental results show that the proposed method is able to generate high-quality layered images and initiates a benchmark for future work.

Abstract: Limited labeled data makes it hard to train models from scratch in medical domain, and an important paradigm is pre-training and then fine-tuning. Large pre-trained models contain rich representations, which can be adapted to downstream medical tasks. However, existing methods either tune all the parameters or the task-specific layers of the pre-trained models, ignoring the input variations of medical images, and thus they are not efficient or effective. In this work, we aim to study parameter-efficient fine-tuning (PEFT) for medical image analysis, and propose a dynamic visual prompt tuning method, named DVPT. It can extract knowledge beneficial to downstream tasks from large models with a few trainable parameters. Firstly, the frozen features are transformed by an lightweight bottleneck layer to learn the domain-specific distribution of downstream medical tasks, and then a few learnable visual prompts are used as dynamic queries and then conduct cross-attention with the transformed features, attempting to acquire sample-specific knowledge that are suitable for each sample. Finally, the features are projected to original feature dimension and aggregated with the frozen features. This DVPT module can be shared between different Transformer layers, further reducing the trainable parameters. To validate DVPT, we conduct extensive experiments with different pre-trained models on medical classification and segmentation tasks. We find such PEFT method can not only efficiently adapt the pre-trained models to the medical domain, but also brings data efficiency with partial labeled data. For example, with 0.5\% extra trainable parameters, our method not only outperforms state-of-the-art PEFT methods, even surpasses the full fine-tuning by more than 2.20\% Kappa score on medical classification task. It can saves up to 60\% labeled data and 99\% storage cost of ViT-B/16.

Abstract: Registration of distant outdoor LiDAR point clouds is crucial to extending the 3D vision of collaborative autonomous vehicles, and yet is challenging due to small overlapping area and a huge disparity between observed point densities. In this paper, we propose Group-wise Contrastive Learning (GCL) scheme to extract density-invariant geometric features to register distant outdoor LiDAR point clouds. We mark through theoretical analysis and experiments that, contrastive positives should be independent and identically distributed (i.i.d.), in order to train densityinvariant feature extractors. We propose upon the conclusion a simple yet effective training scheme to force the feature of multiple point clouds in the same spatial location (referred to as positive groups) to be similar, which naturally avoids the sampling bias introduced by a pair of point clouds to conform with the i.i.d. principle. The resulting fully-convolutional feature extractor is more powerful and density-invariant than state-of-the-art methods, improving the registration recall of distant scenarios on KITTI and nuScenes benchmarks by 40.9% and 26.9%, respectively. The code will be open-sourced.

Abstract: Convolutional neural networks encode images through a sequence of convolutions, normalizations and non-linearities as well as downsampling operations into potentially strong semantic embeddings. Yet, previous work showed that even slight mistakes during sampling, leading to aliasing, can be directly attributed to the networks' lack in robustness. To address such issues and facilitate simpler and faster adversarial training, [12] recently proposed FLC pooling, a method for provably alias-free downsampling - in theory. In this work, we conduct a further analysis through the lens of signal processing and find that such current pooling methods, which address aliasing in the frequency domain, are still prone to spectral leakage artifacts. Hence, we propose aliasing and spectral artifact-free pooling, short ASAP. While only introducing a few modifications to FLC pooling, networks using ASAP as downsampling method exhibit higher native robustness against common corruptions, a property that FLC pooling was missing. ASAP also increases native robustness against adversarial attacks on high and low resolution data while maintaining similar clean accuracy or even outperforming the baseline.

Abstract: Web image datasets curated online inherently contain ambiguous in-distribution (ID) instances and out-of-distribution (OOD) instances, which we collectively call non-conforming (NC) instances. In many recent approaches for mitigating the negative effects of NC instances, the core implicit assumption is that the NC instances can be found via entropy maximization. For "entropy" to be well-defined, we are interpreting the output prediction vector of an instance as the parameter vector of a multinomial random variable, with respect to some trained model with a softmax output layer. Hence, entropy maximization is based on the idealized assumption that NC instances have predictions that are "almost" uniformly distributed. However, in real-world web image datasets, there are numerous NC instances whose predictions are far from being uniformly distributed. To tackle the limitation of entropy maximization, we propose $(\alpha, \beta)$-generalized KL divergence, $\mathcal{D}_{\text{KL}}^{\alpha, \beta}(p\|q)$, which can be used to identify significantly more NC instances. Theoretical properties of $\mathcal{D}_{\text{KL}}^{\alpha, \beta}(p\|q)$ are proven, and we also show empirically that a simple use of $\mathcal{D}_{\text{KL}}^{\alpha, \beta}(p\|q)$ outperforms all baselines on the NC instance identification task. Building upon $(\alpha,\beta)$-generalized KL divergence, we also introduce a new iterative training framework, GenKL, that identifies and relabels NC instances. When evaluated on three web image datasets, Clothing1M, Food101/Food101N, and mini WebVision 1.0, we achieved new state-of-the-art classification accuracies: $81.34\%$, $85.73\%$ and $78.99\%$/$92.54\%$ (top-1/top-5), respectively.

Abstract: Recovering sharp images from dual-pixel (DP) pairs with disparity-dependent blur is a challenging task.~Existing blur map-based deblurring methods have demonstrated promising results. In this paper, we propose, to the best of our knowledge, the first framework to introduce the contrastive language-image pre-training framework (CLIP) to achieve accurate blur map estimation from DP pairs unsupervisedly. To this end, we first carefully design text prompts to enable CLIP to understand blur-related geometric prior knowledge from the DP pair. Then, we propose a format to input stereo DP pair to the CLIP without any fine-tuning, where the CLIP is pre-trained on monocular images. Given the estimated blur map, we introduce a blur-prior attention block, a blur-weighting loss and a blur-aware loss to recover the all-in-focus image. Our method achieves state-of-the-art performance in extensive experiments.

Abstract: In dyadic speaker-listener interactions, the listener's head reactions along with the speaker's head movements, constitute an important non-verbal semantic expression together. The listener Head generation task aims to synthesize responsive listener's head videos based on audios of the speaker and reference images of the listener. Compared to the Talking-head generation, it is more challenging to capture the correlation clues from the speaker's audio and visual information. Following the ViCo baseline scheme, we propose a high-performance solution by enhancing the hierarchical semantic extraction capability of the audio encoder module and improving the decoder part, renderer and post-processing modules. Our solution gets the first place on the official leaderboard for the track of listening head generation. This paper is a technical report of ViCo@2023 Conversational Head Generation Challenge in ACM Multimedia 2023 conference.

Abstract: Despite the wide variety of methods developed for synthetic image attribution, most of them can only attribute images generated by models or architectures included in the training set and do not work with unknown architectures, hindering their applicability in real-world scenarios. In this paper, we propose a verification framework that relies on a Siamese Network to address the problem of open-set attribution of synthetic images to the architecture that generated them. We consider two different settings. In the first setting, the system determines whether two images have been produced by the same generative architecture or not. In the second setting, the system verifies a claim about the architecture used to generate a synthetic image, utilizing one or multiple reference images generated by the claimed architecture. The main strength of the proposed system is its ability to operate in both closed and open-set scenarios so that the input images, either the query and reference images, can belong to the architectures considered during training or not. Experimental evaluations encompassing various generative architectures such as GANs, diffusion models, and transformers, focusing on synthetic face image generation, confirm the excellent performance of our method in both closed and open-set settings, as well as its strong generalization capabilities.

Abstract: Gait recognition is a biometric technique that identifies individuals by their unique walking styles, which is suitable for unconstrained environments and has a wide range of applications. While current methods focus on exploiting body part-based representations, they often neglect the hierarchical dependencies between local motion patterns. In this paper, we propose a hierarchical spatio-temporal representation learning (HSTL) framework for extracting gait features from coarse to fine. Our framework starts with a hierarchical clustering analysis to recover multi-level body structures from the whole body to local details. Next, an adaptive region-based motion extractor (ARME) is designed to learn region-independent motion features. The proposed HSTL then stacks multiple ARMEs in a top-down manner, with each ARME corresponding to a specific partition level of the hierarchy. An adaptive spatio-temporal pooling (ASTP) module is used to capture gait features at different levels of detail to perform hierarchical feature mapping. Finally, a frame-level temporal aggregation (FTA) module is employed to reduce redundant information in gait sequences through multi-scale temporal downsampling. Extensive experiments on CASIA-B, OUMVLP, GREW, and Gait3D datasets demonstrate that our method outperforms the state-of-the-art while maintaining a reasonable balance between model accuracy and complexity.

Abstract: BIQA (Blind Image Quality Assessment) is an important field of study that evaluates images automatically. Although significant progress has been made, blind image quality assessment remains a difficult task since images vary in content and distortions. Most algorithms generate quality without emphasizing the important region of interest. In order to solve this, a multi-stream spatial and channel attention-based algorithm is being proposed. This algorithm generates more accurate predictions with a high correlation to human perceptual assessment by combining hybrid features from two different backbones, followed by spatial and channel attention to provide high weights to the region of interest. Four legacy image quality assessment datasets are used to validate the effectiveness of our proposed approach. Authentic and synthetic distortion image databases are used to demonstrate the effectiveness of the proposed method, and we show that it has excellent generalization properties with a particular focus on the perceptual foreground information.

Abstract: Hyperspectral anomaly detection (HAD) is widely used in Earth observation and deep space exploration. A major challenge for HAD is the complex background of the input hyperspectral images (HSIs), resulting in anomalies confused in the background. On the other hand, the lack of labeled samples for HSIs leads to poor generalization of existing HAD methods. This paper starts the first attempt to study a new and generalizable background learning problem without labeled samples. We present a novel solution BSDM (background suppression diffusion model) for HAD, which can simultaneously learn latent background distributions and generalize to different datasets for suppressing complex background. It is featured in three aspects: (1) For the complex background of HSIs, we design pseudo background noise and learn the potential background distribution in it with a diffusion model (DM). (2) For the generalizability problem, we apply a statistical offset module so that the BSDM adapts to datasets of different domains without labeling samples. (3) For achieving background suppression, we innovatively improve the inference process of DM by feeding the original HSIs into the denoising network, which removes the background as noise. Our work paves a new background suppression way for HAD that can improve HAD performance without the prerequisite of manually labeled data. Assessments and generalization experiments of four HAD methods on several real HSI datasets demonstrate the above three unique properties of the proposed method. The code is available at https://github.com/majitao-xd/BSDM-HAD.

Abstract: Recent advances in machine learning models have greatly increased the performance of automated methods in medical image analysis. However, the internal functioning of such models is largely hidden, which hinders their integration in clinical practice. Explainability and trust are viewed as important aspects of modern methods, for the latter's widespread use in clinical communities. As such, validation of machine learning models represents an important aspect and yet, most methods are only validated in a limited way. In this work, we focus on providing a richer and more appropriate validation approach for highly powerful Visual Question Answering (VQA) algorithms. To better understand the performance of these methods, which answer arbitrary questions related to images, this work focuses on an automatic visual Turing test (VTT). That is, we propose an automatic adaptive questioning method, that aims to expose the reasoning behavior of a VQA algorithm. Specifically, we introduce a reinforcement learning (RL) agent that observes the history of previously asked questions, and uses it to select the next question to pose. We demonstrate our approach in the context of evaluating algorithms that automatically answer questions related to diabetic macular edema (DME) grading. The experiments show that such an agent has similar behavior to a clinician, whereby asking questions that are relevant to key clinical concepts.

Abstract: We propose 3Deformer, a general-purpose framework for interactive 3D shape editing. Given a source 3D mesh with semantic materials, and a user-specified semantic image, 3Deformer can accurately edit the source mesh following the shape guidance of the semantic image, while preserving the source topology as rigid as possible. Recent studies of 3D shape editing mostly focus on learning neural networks to predict 3D shapes, which requires high-cost 3D training datasets and is limited to handling objects involved in the datasets. Unlike these studies, our 3Deformer is a non-training and common framework, which only requires supervision of readily-available semantic images, and is compatible with editing various objects unlimited by datasets. In 3Deformer, the source mesh is deformed utilizing the differentiable renderer technique, according to the correspondences between semantic images and mesh materials. However, guiding complex 3D shapes with a simple 2D image incurs extra challenges, that is, the deform accuracy, surface smoothness, geometric rigidity, and global synchronization of the edited mesh should be guaranteed. To address these challenges, we propose a hierarchical optimization architecture to balance the global and local shape features, and propose further various strategies and losses to improve properties of accuracy, smoothness, rigidity, and so on. Extensive experiments show that our 3Deformer is able to produce impressive results and reaches the state-of-the-art level.

Abstract: Existing 2D-to-3D pose lifting networks suffer from poor performance in cross-dataset benchmarks. Although the use of 2D keypoints joined by "stick-figure" limbs has shown promise as an intermediate step, stick-figures do not account for occlusion information that is often inherent in an image. In this paper, we propose a novel representation using opaque 3D limbs that preserves occlusion information while implicitly encoding joint locations. Crucially, when training on data with accurate three-dimensional keypoints and without part-maps, this representation allows training on abstract synthetic images, with occlusion, from as many synthetic viewpoints as desired. The result is a pose defined by limb angles rather than joint positions $\unicode{x2013}$ because poses are, in the real world, independent of cameras $\unicode{x2013}$ allowing us to predict poses that are completely independent of camera viewpoint. The result provides not only an improvement in same-dataset benchmarks, but a "quantum leap" in cross-dataset benchmarks.

Abstract: Talking head video generation aims to animate a human face in a still image with dynamic poses and expressions using motion information derived from a target-driving video, while maintaining the person's identity in the source image. However, dramatic and complex motions in the driving video cause ambiguous generation, because the still source image cannot provide sufficient appearance information for occluded regions or delicate expression variations, which produces severe artifacts and significantly degrades the generation quality. To tackle this problem, we propose to learn a global facial representation space, and design a novel implicit identity representation conditioned memory compensation network, coined as MCNet, for high-fidelity talking head generation.~Specifically, we devise a network module to learn a unified spatial facial meta-memory bank from all training samples, which can provide rich facial structure and appearance priors to compensate warped source facial features for the generation. Furthermore, we propose an effective query mechanism based on implicit identity representations learned from the discrete keypoints of the source image. It can greatly facilitate the retrieval of more correlated information from the memory bank for the compensation. Extensive experiments demonstrate that MCNet can learn representative and complementary facial memory, and can clearly outperform previous state-of-the-art talking head generation methods on VoxCeleb1 and CelebV datasets. Please check our \href{https://github.com/harlanhong/ICCV2023-MCNET}{Project}.

Abstract: Cross-lingual image captioning is confronted with both cross-lingual and cross-modal challenges for multimedia analysis. The crucial issue in this task is to model the global and local matching between the image and different languages. Existing cross-modal embedding methods based on Transformer architecture oversight the local matching between the image region and monolingual words, not to mention in the face of a variety of differentiated languages. Due to the heterogeneous property of the cross-modal and cross-lingual task, we utilize the heterogeneous network to establish cross-domain relationships and the local correspondences between the image and different languages. In this paper, we propose an Embedded Heterogeneous Attention Transformer (EHAT) to build reasoning paths bridging cross-domain for cross-lingual image captioning and integrate into transformer. The proposed EHAT consists of a Masked Heterogeneous Cross-attention (MHCA), Heterogeneous Attention Reasoning Network (HARN) and Heterogeneous Co-attention (HCA). HARN as the core network, models and infers cross-domain relationship anchored by vision bounding box representation features to connect two languages word features and learn the heterogeneous maps. MHCA and HCA implement cross-domain integration in the encoder through the special heterogeneous attention and enable single model to generate two language captioning. We test on MSCOCO dataset to generate English and Chinese, which are most widely used and have obvious difference between their language families. Our experiments show that our method even achieve better than advanced monolingual methods.

Abstract: Effectively measuring and modeling the reliability of a trained model is essential to the real-world deployment of monocular depth estimation (MDE) models. However, the intrinsic ill-posedness and ordinal-sensitive nature of MDE pose major challenges to the estimation of uncertainty degree of the trained models. On the one hand, utilizing current uncertainty modeling methods may increase memory consumption and are usually time-consuming. On the other hand, measuring the uncertainty based on model accuracy can also be problematic, where uncertainty reliability and prediction accuracy are not well decoupled. In this paper, we propose to model the uncertainty of MDE models from the perspective of the inherent probability distributions originating from the depth probability volume and its extensions, and to assess it more fairly with more comprehensive metrics. By simply introducing additional training regularization terms, our model, with surprisingly simple formations and without requiring extra modules or multiple inferences, can provide uncertainty estimations with state-of-the-art reliability, and can be further improved when combined with ensemble or sampling methods. A series of experiments demonstrate the effectiveness of our methods.

Abstract: Multimodal image registration is a challenging but essential step for numerous image-guided procedures. Most registration algorithms rely on the computation of complex, frequently non-differentiable similarity metrics to deal with the appearance discrepancy of anatomical structures between imaging modalities. Recent Machine Learning based approaches are limited to specific anatomy-modality combinations and do not generalize to new settings. We propose a generic framework for creating expressive cross-modal descriptors that enable fast deformable global registration. We achieve this by approximating existing metrics with a dot-product in the feature space of a small convolutional neural network (CNN) which is inherently differentiable can be trained without registered data. Our method is several orders of magnitude faster than local patch-based metrics and can be directly applied in clinical settings by replacing the similarity measure with the proposed one. Experiments on three different datasets demonstrate that our approach generalizes well beyond the training data, yielding a broad capture range even on unseen anatomies and modality pairs, without the need for specialized retraining. We make our training code and data publicly available.

Abstract: This paper focuses on motion prediction for point cloud sequences in the challenging case of deformable 3D objects, such as human body motion. First, we investigate the challenges caused by deformable shapes and complex motions present in this type of representation, with the ultimate goal of understanding the technical limitations of state-of-the-art models. From this understanding, we propose an improved architecture for point cloud prediction of deformable 3D objects. Specifically, to handle deformable shapes, we propose a graph-based approach that learns and exploits the spatial structure of point clouds to extract more representative features. Then we propose a module able to combine the learned features in an adaptative manner according to the point cloud movements. The proposed adaptative module controls the composition of local and global motions for each point, enabling the network to model complex motions in deformable 3D objects more effectively. We tested the proposed method on the following datasets: MNIST moving digits, the Mixamo human bodies motions, JPEG and CWIPC-SXR real-world dynamic bodies. Simulation results demonstrate that our method outperforms the current baseline methods given its improved ability to model complex movements as well as preserve point cloud shape. Furthermore, we demonstrate the generalizability of the proposed framework for dynamic feature learning, by testing the framework for action recognition on the MSRAction3D dataset and achieving results on-par with state-of-the-art methods

Abstract: Capsule Networks have emerged as a powerful class of deep learning architectures, known for robust performance with relatively few parameters compared to Convolutional Neural Networks (CNNs). However, their inherent efficiency is often overshadowed by their slow, iterative routing mechanisms which establish connections between Capsule layers, posing computational challenges resulting in an inability to scale. In this paper, we introduce a novel, non-iterative routing mechanism, inspired by trainable prototype clustering. This innovative approach aims to mitigate computational complexity, while retaining, if not enhancing, performance efficacy. Furthermore, we harness a shared Capsule subspace, negating the need to project each lower-level Capsule to each higher-level Capsule, thereby significantly reducing memory requisites during training. Our approach demonstrates superior results compared to the current best non-iterative Capsule Network and tests on the Imagewoof dataset, which is too computationally demanding to handle efficiently by iterative approaches. Our findings underscore the potential of our proposed methodology in enhancing the operational efficiency and performance of Capsule Networks, paving the way for their application in increasingly complex computational scenarios.

Abstract: Deep neural networks have shown exceptional performance in various tasks, but their lack of robustness, reliability, and tendency to be overconfident pose challenges for their deployment in safety-critical applications like autonomous driving. In this regard, quantifying the uncertainty inherent to a model's prediction is a promising endeavour to address these shortcomings. In this work, we present a novel Uncertainty-aware Cross-Entropy loss (U-CE) that incorporates dynamic predictive uncertainties into the training process by pixel-wise weighting of the well-known cross-entropy loss (CE). Through extensive experimentation, we demonstrate the superiority of U-CE over regular CE training on two benchmark datasets, Cityscapes and ACDC, using two common backbone architectures, ResNet-18 and ResNet-101. With U-CE, we manage to train models that not only improve their segmentation performance but also provide meaningful uncertainties after training. Consequently, we contribute to the development of more robust and reliable segmentation models, ultimately advancing the state-of-the-art in safety-critical applications and beyond.

Abstract: Visual (re)localization is critical for various applications in computer vision and robotics. Its goal is to estimate the 6 degrees of freedom (DoF) camera pose for each query image, based on a set of posed database images. Currently, all leading solutions are structure-based that either explicitly construct 3D metric maps from the database with structure-from-motion, or implicitly encode the 3D information with scene coordinate regression models. On the contrary, visual localization without reconstructing the scene in 3D offers clear benefits. It makes deployment more convenient by reducing database pre-processing time, releasing storage requirements, and remaining unaffected by imperfect reconstruction, etc. In this technical report, we demonstrate that it is possible to achieve high localization accuracy without reconstructing the scene from the database. The key to achieving this owes to a tailored motion averaging over database-query pairs. Experiments show that our visual localization proposal, LazyLoc, achieves comparable performance against state-of-the-art structure-based methods. Furthermore, we showcase the versatility of LazyLoc, which can be easily extended to handle complex configurations such as multi-query co-localization and camera rigs.

Abstract: To enable context-aware computer assistance in the operating room of the future, cognitive systems need to understand automatically which surgical phase is being performed by the medical team. The primary source of information for surgical phase recognition is typically video, which presents two challenges: extracting meaningful features from the video stream and effectively modeling temporal information in the sequence of visual features. For temporal modeling, attention mechanisms have gained popularity due to their ability to capture long-range dependencies. In this paper, we explore design choices for attention in existing temporal models for surgical phase recognition and propose a novel approach that does not resort to local attention or regularization of attention weights: TUNeS is an efficient and simple temporal model that incorporates self-attention at the coarsest stage of a U-Net-like structure. In addition, we propose to train the feature extractor, a standard CNN, together with an LSTM on preferably long video segments, i.e., with long temporal context. In our experiments, all temporal models performed better on top of feature extractors that were trained with longer temporal context. On top of these contextualized features, TUNeS achieves state-of-the-art results on Cholec80.

Abstract: Motivated by the recent trend towards the usage of larger receptive fields for more context-aware neural networks in vision applications, we aim to investigate how large these receptive fields really need to be. To facilitate such study, several challenges need to be addressed, most importantly: (i) We need to provide an effective way for models to learn large filters (potentially as large as the input data) without increasing their memory consumption during training or inference, (ii) the study of filter sizes has to be decoupled from other effects such as the network width or number of learnable parameters, and (iii) the employed convolution operation should be a plug-and-play module that can replace any conventional convolution in a Convolutional Neural Network (CNN) and allow for an efficient implementation in current frameworks. To facilitate such models, we propose to learn not spatial but frequency representations of filter weights as neural implicit functions, such that even infinitely large filters can be parameterized by only a few learnable weights. The resulting neural implicit frequency CNNs are the first models to achieve results on par with the state-of-the-art on large image classification benchmarks while executing convolutions solely in the frequency domain and can be employed within any CNN architecture. They allow us to provide an extensive analysis of the learned receptive fields. Interestingly, our analysis shows that, although the proposed networks could learn very large convolution kernels, the learned filters practically translate into well-localized and relatively small convolution kernels in the spatial domain.

Abstract: The field of Explainable Artificial Intelligence (XAI) aims to improve the interpretability of black-box machine learning models. Building a heatmap based on the importance value of input features is a popular method for explaining the underlying functions of such models in producing their predictions. Heatmaps are almost understandable to humans, yet they are not without flaws. Non-expert users, for example, may not fully understand the logic of heatmaps (the logic in which relevant pixels to the model's prediction are highlighted with different intensities or colors). Additionally, objects and regions of the input image that are relevant to the model prediction are frequently not entirely differentiated by heatmaps. In this paper, we propose a framework called TbExplain that employs XAI techniques and a pre-trained object detector to present text-based explanations of scene classification models. Moreover, TbExplain incorporates a novel method to correct predictions and textually explain them based on the statistics of objects in the input image when the initial prediction is unreliable. To assess the trustworthiness and validity of the text-based explanations, we conducted a qualitative experiment, and the findings indicated that these explanations are sufficiently reliable. Furthermore, our quantitative and qualitative experiments on TbExplain with scene classification datasets reveal an improvement in classification accuracy over ResNet variants.

Abstract: Audio-driven portrait animation aims to synthesize portrait videos that are conditioned by given audio. Animating high-fidelity and multimodal video portraits has a variety of applications. Previous methods have attempted to capture different motion modes and generate high-fidelity portrait videos by training different models or sampling signals from given videos. However, lacking correlation learning between lip-sync and other movements (e.g., head pose/eye blinking) usually leads to unnatural results. In this paper, we propose a unified system for multi-person, diverse, and high-fidelity talking portrait generation. Our method contains three stages, i.e., 1) Mapping-Once network with Dual Attentions (MODA) generates talking representation from given audio. In MODA, we design a dual-attention module to encode accurate mouth movements and diverse modalities. 2) Facial composer network generates dense and detailed face landmarks, and 3) temporal-guided renderer syntheses stable videos. Extensive evaluations demonstrate that the proposed system produces more natural and realistic video portraits compared to previous methods.

Abstract: Deep learning-based person identification and verification systems have remarkably improved in terms of accuracy in recent years; however, such systems, including widely popular cloud-based solutions, have been found to exhibit significant biases related to race, age, and gender, a problem that requires in-depth exploration and solutions. This paper presents an in-depth analysis, with a particular emphasis on the intersectionality of these demographic factors. Intersectional bias refers to the performance discrepancies w.r.t. the different combinations of race, age, and gender groups, an area relatively unexplored in current literature. Furthermore, the reliance of most state-of-the-art approaches on accuracy as the principal evaluation metric often masks significant demographic disparities in performance. To counter this crucial limitation, we incorporate five additional metrics in our quantitative analysis, including disparate impact and mistreatment metrics, which are typically ignored by the relevant fairness-aware approaches. Results on the Racial Faces in-the-Wild (RFW) benchmark indicate pervasive biases in face recognition systems, extending beyond race, with different demographic factors yielding significantly disparate outcomes. In particular, Africans demonstrate an 11.25% lower True Positive Rate (TPR) compared to Caucasians, while only a 3.51% accuracy drop is observed. Even more concerning, the intersections of multiple protected groups, such as African females over 60 years old, demonstrate a +39.89% disparate mistreatment rate compared to the highest Caucasians rate. By shedding light on these biases and their implications, this paper aims to stimulate further research towards developing fairer, more equitable face recognition and verification systems.

Abstract: Automated medical image classification is the key component in intelligent diagnosis systems. However, most medical image datasets contain plenty of samples of common diseases and just a handful of rare ones, leading to major class imbalances. Currently, it is an open problem in intelligent diagnosis to effectively learn from imbalanced training data. In this paper, we propose a simple yet effective framework, named \textbf{C}lass \textbf{A}ttention to \textbf{RE}gions of the lesion (CARE), to handle data imbalance issues by embedding attention into the training process of \textbf{C}onvolutional \textbf{N}eural \textbf{N}etworks (CNNs). The proposed attention module helps CNNs attend to lesion regions of rare diseases, therefore helping CNNs to learn their characteristics more effectively. In addition, this attention module works only during the training phase and does not change the architecture of the original network, so it can be directly combined with any existing CNN architecture. The CARE framework needs bounding boxes to represent the lesion regions of rare diseases. To alleviate the need for manual annotation, we further developed variants of CARE by leveraging the traditional saliency methods or a pretrained segmentation model for bounding box generation. Results show that the CARE variants with automated bounding box generation are comparable to the original CARE framework with \textit{manual} bounding box annotations. A series of experiments on an imbalanced skin image dataset and a pneumonia dataset indicates that our method can effectively help the network focus on the lesion regions of rare diseases and remarkably improves the classification performance of rare diseases.

Abstract: Multimodal vision-language (VL) learning has noticeably pushed the tendency toward generic intelligence owing to emerging large foundation models. However, tracking, as a fundamental vision problem, surprisingly enjoys less bonus from recent flourishing VL learning. We argue that the reasons are two-fold: the lack of large-scale vision-language annotated videos and ineffective vision-language interaction learning of current works. These nuisances motivate us to design more effective vision-language representation for tracking, meanwhile constructing a large database with language annotation for model learning. Particularly, in this paper, we first propose a general attribute annotation strategy to decorate videos in six popular tracking benchmarks, which contributes a large-scale vision-language tracking database with more than 23,000 videos. We then introduce a novel framework to improve tracking by learning a unified-adaptive VL representation, where the cores are the proposed asymmetric architecture search and modality mixer (ModaMixer). To further improve VL representation, we introduce a contrastive loss to align different modalities. To thoroughly evidence the effectiveness of our method, we integrate the proposed framework on three tracking methods with different designs, i.e., the CNN-based SiamCAR, the Transformer-based OSTrack, and the hybrid structure TransT. The experiments demonstrate that our framework can significantly improve all baselines on six benchmarks. Besides empirical results, we theoretically analyze our approach to show its rationality. By revealing the potential of VL representation, we expect the community to divert more attention to VL tracking and hope to open more possibilities for future tracking with diversified multimodal messages.

Abstract: Deploying deep visual models can lead to performance drops due to the discrepancies between source and target distributions. Several approaches leverage labeled source data to estimate target domain accuracy, but accessing labeled source data is often prohibitively difficult due to data confidentiality or resource limitations on serving devices. Our work proposes a new framework to estimate model accuracy on unlabeled target data without access to source data. We investigate the feasibility of using pseudo-labels for accuracy estimation and evolve this idea into adopting recent advances in source-free domain adaptation algorithms. Our approach measures the disagreement rate between the source hypothesis and the target pseudo-labeling function, adapted from the source hypothesis. We mitigate the impact of erroneous pseudo-labels that may arise due to a high ideal joint hypothesis risk by employing adaptive adversarial perturbation on the input of the target model. Our proposed source-free framework effectively addresses the challenging distribution shift scenarios and outperforms existing methods requiring source data and labels for training.

Abstract: Prototype-based classification is a classical method in machine learning, and recently it has achieved remarkable success in semi-supervised semantic segmentation. However, the current approach isolates the prototype initialization process from the main training framework, which appears to be unnecessary. Furthermore, while the direct use of K-Means algorithm for prototype generation has considered rich intra-class variance, it may not be the optimal solution for the classification task. To tackle these problems, we propose a novel boundary-refined prototype generation (BRPG) method, which is incorporated into the whole training framework. Specifically, our approach samples and clusters high- and low-confidence features separately based on a confidence threshold, aiming to generate prototypes closer to the class boundaries. Moreover, an adaptive prototype optimization strategy is introduced to make prototype augmentation for categories with scattered feature distributions. Extensive experiments on the PASCAL VOC 2012 and Cityscapes datasets demonstrate the superiority and scalability of the proposed method, outperforming the current state-of-the-art approaches. The code is available at xxxxxxxxxxxxxx.

Abstract: Though performed almost effortlessly by humans, segmenting 2D gray-scale or color images in terms of their constituent regions of interest (e.g.~background, objects or portions of objects) constitutes one of the greatest challenges in science and technology as a consequence of the involved dimensionality reduction(3D to 2D), noise, reflections, shades, and occlusions, among many other possible effects. While a large number of interesting approaches have been respectively suggested along the last decades, it was mainly with the more recent development of deep learning that more effective and general solutions have been obtained, currently constituting the basic comparison reference for this type of operation. Also developed recently, a multiset-based methodology has been described that is capable of encouraging performance that combines spatial accuracy, stability, and robustness while requiring minimal computational resources (hardware and/or training and recognition time). The interesting features of the latter methodology mostly follow from the enhanced selectivity and sensitivity, as well as good robustness to data perturbations and outliers, allowed by the coincidence similarity index on which the multiset approach to supervised image segmentation is based. After describing the deep learning and multiset approaches, the present work develops two comparison experiments between them which are primarily aimed at illustrating their respective main interesting features when applied to the adopted specific type of data and parameter configurations. While the deep learning approach confirmed its potential for performing image segmentation, the alternative multiset methodology allowed for encouraging accuracy while requiring little computational resources.

Abstract: Facial age estimation has received a lot of attention for its diverse application scenarios. Most existing studies treat each sample equally and aim to reduce the average estimation error for the entire dataset, which can be summarized as General Age Estimation. However, due to the long-tailed distribution prevalent in the dataset, treating all samples equally will inevitably bias the model toward the head classes (usually the adult with a majority of samples). Driven by this, some works suggest that each class should be treated equally to improve performance in tail classes (with a minority of samples), which can be summarized as Long-tailed Age Estimation. However, Long-tailed Age Estimation usually faces a performance trade-off, i.e., achieving improvement in tail classes by sacrificing the head classes. In this paper, our goal is to design a unified framework to perform well on both tasks, killing two birds with one stone. To this end, we propose a simple, effective, and flexible training paradigm named GLAE, which is two-fold. Our GLAE provides a surprising improvement on Morph II, reaching the lowest MAE and CMAE of 1.14 and 1.27 years, respectively. Compared to the previous best method, MAE dropped by up to 34%, which is an unprecedented improvement, and for the first time, MAE is close to 1 year old. Extensive experiments on other age benchmark datasets, including CACD, MIVIA, and Chalearn LAP 2015, also indicate that GLAE outperforms the state-of-the-art approaches significantly.

Abstract: Lip reading is a challenging task that has many potential applications in speech recognition, human-computer interaction, and security systems. However, existing lip reading systems often suffer from low accuracy due to the limitations of video features. In this paper, we propose a novel approach that leverages visemes, which are groups of phonetically similar lip shapes, to extract more discriminative and robust video features for lip reading. We evaluate our approach on various tasks, including word-level and sentence-level lip reading, and audiovisual speech recognition using the Arman-AV dataset, a largescale Persian corpus. Our experimental results show that our viseme based approach consistently outperforms the state-of-theart methods in all these tasks. The proposed method reduces the lip-reading word error rate (WER) by 9.1% relative to the best previous method.

Abstract: In an era where visual content generation is increasingly driven by machine learning, the integration of human feedback into generative models presents significant opportunities for enhancing user experience and output quality. This study explores strategies for incorporating iterative human feedback into the generative process of diffusion-based text-to-image models. We propose FABRIC, a training-free approach applicable to a wide range of popular diffusion models, which exploits the self-attention layer present in the most widely used architectures to condition the diffusion process on a set of feedback images. To ensure a rigorous assessment of our approach, we introduce a comprehensive evaluation methodology, offering a robust mechanism to quantify the performance of generative visual models that integrate human feedback. We show that generation results improve over multiple rounds of iterative feedback through exhaustive analysis, implicitly optimizing arbitrary user preferences. The potential applications of these findings extend to fields such as personalized content creation and customization.

Abstract: Millions of vehicles are transported every year, tightly parked in vessels or boats. To reduce the risks of associated safety issues like fires, knowing the location of vehicles is essential, since different vehicles may need different mitigation measures, e.g. electric cars. This work is aimed at creating a solution based on a nano-drone that navigates across rows of parked vehicles and detects their license plates. We do so via a wall-following algorithm, and a CNN trained to detect license plates. All computations are done in real-time on the drone, which just sends position and detected images that allow the creation of a 2D map with the position of the plates. Our solution is capable of reading all plates across eight test cases (with several rows of plates, different drone speeds, or low light) by aggregation of measurements across several drone journeys.

Abstract: Generative Engineering Design approaches driven by Deep Generative Models (DGM) have been proposed to facilitate industrial engineering processes. In such processes, designs often come in the form of images, such as blueprints, engineering drawings, and CAD models depending on the level of detail. DGMs have been successfully employed for synthesis of natural images, e.g., displaying animals, human faces and landscapes. However, industrial design images are fundamentally different from natural scenes in that they contain rich structural patterns and long-range dependencies, which are challenging for convolution-based DGMs to generate. Moreover, DGM-driven generation process is typically triggered based on random noisy inputs, which outputs unpredictable samples and thus cannot perform an efficient industrial design exploration. We tackle these challenges by proposing a novel model Self-Attention Adversarial Latent Autoencoder (SA-ALAE), which allows generating feasible design images of complex engineering parts. With SA-ALAE, users can not only explore novel variants of an existing design, but also control the generation process by operating in latent space. The potential of SA-ALAE is shown by generating engineering blueprints in a real automotive design task.

Abstract: Realistic human-centric rendering plays a key role in both computer vision and computer graphics. Rapid progress has been made in the algorithm aspect over the years, yet existing human-centric rendering datasets and benchmarks are rather impoverished in terms of diversity, which are crucial for rendering effect. Researchers are usually constrained to explore and evaluate a small set of rendering problems on current datasets, while real-world applications require methods to be robust across different scenarios. In this work, we present DNA-Rendering, a large-scale, high-fidelity repository of human performance data for neural actor rendering. DNA-Rendering presents several alluring attributes. First, our dataset contains over 1500 human subjects, 5000 motion sequences, and 67.5M frames' data volume. Second, we provide rich assets for each subject -- 2D/3D human body keypoints, foreground masks, SMPLX models, cloth/accessory materials, multi-view images, and videos. These assets boost the current method's accuracy on downstream rendering tasks. Third, we construct a professional multi-view system to capture data, which contains 60 synchronous cameras with max 4096 x 3000 resolution, 15 fps speed, and stern camera calibration steps, ensuring high-quality resources for task training and evaluation. Along with the dataset, we provide a large-scale and quantitative benchmark in full-scale, with multiple tasks to evaluate the existing progress of novel view synthesis, novel pose animation synthesis, and novel identity rendering methods. In this manuscript, we describe our DNA-Rendering effort as a revealing of new observations, challenges, and future directions to human-centric rendering. The dataset, code, and benchmarks will be publicly available at https://dna-rendering.github.io/