Search

Linear attention mechanisms have gained prominence in causal language models due to their linear computational complexity and enhanced speed. However, the inherent decay mechanism in linear attention presents challenges when applied to multi-dimensional sequence modeling tasks, such as image processing and multi-modal learning. In these scenarios, the utilization of sequential scanning to establish a global receptive field necessitates multiple scans for multi-dimensional data, thereby leading to inefficiencies. This paper identifies the inefficiency caused by a multiplicative linear recurrence and proposes an efficient alternative additive linear recurrence to avoid the issue, as it can handle multi-dimensional data within a single scan. We further develop an efficient multi-dimensional sequential modeling framework called LightNet based on the new recurrence. Moreover, we present two new multi-dimensional linear relative positional encoding methods, MD-TPE and MD-LRPE to enhance the model's ability to discern positional information in multi-dimensional scenarios. Our empirical evaluations across various tasks, including image classification, image generation, bidirectional language modeling, and autoregressive language modeling, demonstrate the efficacy of LightNet, showcasing its potential as a versatile and efficient solution for multi-dimensional sequential modeling., Comment: Technical report. Yiran Zhong is the corresponding author. The code is available at https://github.com/OpenNLPLab/LightNet

Even though Non-rigid Structure-from-Motion (NRSfM) has been extensively studied and great progress has been made, there are still key challenges that hinder their broad real-world applications: 1) the inherent motion/rotation ambiguity requires either explicit camera motion recovery with extra constraint or complex Procrustean Alignment; 2) existing low-rank modeling of the global shape can over-penalize drastic deformations in the 3D shape sequence. This paper proposes to resolve the above issues from a spatial-temporal modeling perspective. First, we propose a novel Temporally-smooth Procrustean Alignment module that estimates 3D deforming shapes and adjusts the camera motion by aligning the 3D shape sequence consecutively. Our new alignment module remedies the requirement of complex reference 3D shape during alignment, which is more conductive to non-isotropic deformation modeling. Second, we propose a spatial-weighted approach to enforce the low-rank constraint adaptively at different locations to accommodate drastic spatially-variant deformation reconstruction better. Our modeling outperform existing low-rank based methods, and extensive experiments across different datasets validate the effectiveness of our method., Comment: Accepted by CVPR 2024; V2 adds new experiments

The Text to Audible-Video Generation (TAVG) task involves generating videos with accompanying audio based on text descriptions. Achieving this requires skillful alignment of both audio and video elements. To support research in this field, we have developed a comprehensive Text to Audible-Video Generation Benchmark (TAVGBench), which contains over 1.7 million clips with a total duration of 11.8 thousand hours. We propose an automatic annotation pipeline to ensure each audible video has detailed descriptions for both its audio and video contents. We also introduce the Audio-Visual Harmoni score (AVHScore) to provide a quantitative measure of the alignment between the generated audio and video modalities. Additionally, we present a baseline model for TAVG called TAVDiffusion, which uses a two-stream latent diffusion model to provide a fundamental starting point for further research in this area. We achieve the alignment of audio and video by employing cross-attention and contrastive learning. Through extensive experiments and evaluations on TAVGBench, we demonstrate the effectiveness of our proposed model under both conventional metrics and our proposed metrics., Comment: Technical Report. Project page:https://github.com/OpenNLPLab/TAVGBench

In this paper, we propose a 3D geometry-aware deformable Gaussian Splatting method for dynamic view synthesis. Existing neural radiance fields (NeRF) based solutions learn the deformation in an implicit manner, which cannot incorporate 3D scene geometry. Therefore, the learned deformation is not necessarily geometrically coherent, which results in unsatisfactory dynamic view synthesis and 3D dynamic reconstruction. Recently, 3D Gaussian Splatting provides a new representation of the 3D scene, building upon which the 3D geometry could be exploited in learning the complex 3D deformation. Specifically, the scenes are represented as a collection of 3D Gaussian, where each 3D Gaussian is optimized to move and rotate over time to model the deformation. To enforce the 3D scene geometry constraint during deformation, we explicitly extract 3D geometry features and integrate them in learning the 3D deformation. In this way, our solution achieves 3D geometry-aware deformation modeling, which enables improved dynamic view synthesis and 3D dynamic reconstruction. Extensive experimental results on both synthetic and real datasets prove the superiority of our solution, which achieves new state-of-the-art performance. The project is available at https://npucvr.github.io/GaGS/, Comment: Accepted by CVPR 2024. Project page: https://npucvr.github.io/GaGS/

Existing deep learning-based depth completion methods generally employ massive stacked layers to predict the dense depth map from sparse input data. Although such approaches greatly advance this task, their accompanied huge computational complexity hinders their practical applications. To accomplish depth completion more efficiently, we propose a novel lightweight deep network framework, the Long-short Range Recurrent Updating (LRRU) network. Without learning complex feature representations, LRRU first roughly fills the sparse input to obtain an initial dense depth map, and then iteratively updates it through learned spatially-variant kernels. Our iterative update process is content-adaptive and highly flexible, where the kernel weights are learned by jointly considering the guidance RGB images and the depth map to be updated, and large-to-small kernel scopes are dynamically adjusted to capture long-to-short range dependencies. Our initial depth map has coarse but complete scene depth information, which helps relieve the burden of directly regressing the dense depth from sparse ones, while our proposed method can effectively refine it to an accurate depth map with less learnable parameters and inference time. Experimental results demonstrate that our proposed LRRU variants achieve state-of-the-art performance across different parameter regimes. In particular, the LRRU-Base model outperforms competing approaches on the NYUv2 dataset, and ranks 1st on the KITTI depth completion benchmark at the time of submission. Project page: https://npucvr.github.io/LRRU/., Comment: Published in ICCV 2023

We propose an Explicit Conditional Multimodal Variational Auto-Encoder (ECMVAE) for audio-visual segmentation (AVS), aiming to segment sound sources in the video sequence. Existing AVS methods focus on implicit feature fusion strategies, where models are trained to fit the discrete samples in the dataset. With a limited and less diverse dataset, the resulting performance is usually unsatisfactory. In contrast, we address this problem from an effective representation learning perspective, aiming to model the contribution of each modality explicitly. Specifically, we find that audio contains critical category information of the sound producers, and visual data provides candidate sound producer(s). Their shared information corresponds to the target sound producer(s) shown in the visual data. In this case, cross-modal shared representation learning is especially important for AVS. To achieve this, our ECMVAE factorizes the representations of each modality with a modality-shared representation and a modality-specific representation. An orthogonality constraint is applied between the shared and specific representations to maintain the exclusive attribute of the factorized latent code. Further, a mutual information maximization regularizer is introduced to achieve extensive exploration of each modality. Quantitative and qualitative evaluations on the AVSBench demonstrate the effectiveness of our approach, leading to a new state-of-the-art for AVS, with a 3.84 mIOU performance leap on the challenging MS3 subset for multiple sound source segmentation., Comment: Accepted by ICCV2023,Project page(https://npucvr.github.io/MMVAE-AVS),Code(https://github.com/OpenNLPLab/MMVAE-AVS)

This paper proposes a neural radiance field (NeRF) approach for novel view synthesis of dynamic scenes using forward warping. Existing methods often adopt a static NeRF to represent the canonical space, and render dynamic images at other time steps by mapping the sampled 3D points back to the canonical space with the learned backward flow field. However, this backward flow field is non-smooth and discontinuous, which is difficult to be fitted by commonly used smooth motion models. To address this problem, we propose to estimate the forward flow field and directly warp the canonical radiance field to other time steps. Such forward flow field is smooth and continuous within the object region, which benefits the motion model learning. To achieve this goal, we represent the canonical radiance field with voxel grids to enable efficient forward warping, and propose a differentiable warping process, including an average splatting operation and an inpaint network, to resolve the many-to-one and one-to-many mapping issues. Thorough experiments show that our method outperforms existing methods in both novel view rendering and motion modeling, demonstrating the effectiveness of our forward flow motion modeling. Project page: https://npucvr.github.io/ForwardFlowDNeRF, Comment: Accepted by ICCV2023 as oral. Project page: https://npucvr.github.io/ForwardFlowDNeRF

Recently, the RGB images and point clouds fusion methods have been proposed to jointly estimate 2D optical flow and 3D scene flow. However, as both conventional RGB cameras and LiDAR sensors adopt a frame-based data acquisition mechanism, their performance is limited by the fixed low sampling rates, especially in highly-dynamic scenes. By contrast, the event camera can asynchronously capture the intensity changes with a very high temporal resolution, providing complementary dynamic information of the observed scenes. In this paper, we incorporate RGB images, Point clouds and Events for joint optical flow and scene flow estimation with our proposed multi-stage multimodal fusion model, RPEFlow. First, we present an attention fusion module with a cross-attention mechanism to implicitly explore the internal cross-modal correlation for 2D and 3D branches, respectively. Second, we introduce a mutual information regularization term to explicitly model the complementary information of three modalities for effective multimodal feature learning. We also contribute a new synthetic dataset to advocate further research. Experiments on both synthetic and real datasets show that our model outperforms the existing state-of-the-art by a wide margin. Code and dataset is available at https://npucvr.github.io/RPEFlow., Comment: ICCV 2023. Project page: https://npucvr.github.io/RPEFlow Code: https://github.com/danqu130/RPEFlow

RGB-guided depth completion aims at predicting dense depth maps from sparse depth measurements and corresponding RGB images, where how to effectively and efficiently exploit the multi-modal information is a key issue. Guided dynamic filters, which generate spatially-variant depth-wise separable convolutional filters from RGB features to guide depth features, have been proven to be effective in this task. However, the dynamically generated filters require massive model parameters, computational costs and memory footprints when the number of feature channels is large. In this paper, we propose to decompose the guided dynamic filters into a spatially-shared component multiplied by content-adaptive adaptors at each spatial location. Based on the proposed idea, we introduce two decomposition schemes A and B, which decompose the filters by splitting the filter structure and using spatial-wise attention, respectively. The decomposed filters not only maintain the favorable properties of guided dynamic filters as being content-dependent and spatially-variant, but also reduce model parameters and hardware costs, as the learned adaptors are decoupled with the number of feature channels. Extensive experimental results demonstrate that the methods using our schemes outperform state-of-the-art methods on the KITTI dataset, and rank 1st and 2nd on the KITTI benchmark at the time of submission. Meanwhile, they also achieve comparable performance on the NYUv2 dataset. In addition, our proposed methods are general and could be employed as plug-and-play feature fusion blocks in other multi-modal fusion tasks such as RGB-D salient object detection.

Most of the previous 3D human pose estimation work relied on the powerful memory capability of the network to obtain suitable 2D-3D mappings from the training data. Few works have studied the modeling of human posture deformation in motion. In this paper, we propose a new modeling method for human pose deformations and design an accompanying diffusion-based motion prior. Inspired by the field of non-rigid structure-from-motion, we divide the task of reconstructing 3D human skeletons in motion into the estimation of a 3D reference skeleton, and a frame-by-frame skeleton deformation. A mixed spatial-temporal NRSfMformer is used to simultaneously estimate the 3D reference skeleton and the skeleton deformation of each frame from 2D observations sequence, and then sum them to obtain the pose of each frame. Subsequently, a loss term based on the diffusion model is used to ensure that the pipeline learns the correct prior motion knowledge. Finally, we have evaluated our proposed method on mainstream datasets and obtained superior results outperforming the state-of-the-art.

The aim of audio-visual segmentation (AVS) is to precisely differentiate audible objects within videos down to the pixel level. Traditional approaches often tackle this challenge by combining information from various modalities, where the contribution of each modality is implicitly or explicitly modeled. Nevertheless, the interconnections between different modalities tend to be overlooked in audio-visual modeling. In this paper, inspired by the human ability to mentally simulate the sound of an object and its visual appearance, we introduce a bidirectional generation framework. This framework establishes robust correlations between an object's visual characteristics and its associated sound, thereby enhancing the performance of AVS. To achieve this, we employ a visual-to-audio projection component that reconstructs audio features from object segmentation masks and minimizes reconstruction errors. Moreover, recognizing that many sounds are linked to object movements, we introduce an implicit volumetric motion estimation module to handle temporal dynamics that may be challenging to capture using conventional optical flow methods. To showcase the effectiveness of our approach, we conduct comprehensive experiments and analyses on the widely recognized AVSBench benchmark. As a result, we establish a new state-of-the-art performance level in the AVS benchmark, particularly excelling in the challenging MS3 subset which involves segmenting multiple sound sources. To facilitate reproducibility, we plan to release both the source code and the pre-trained model., Comment: AAAI Camera Ready. Dawei Hao and Yuxin Mao contribute equality to this paper. Yiran Zhong is the corresponding author. The code will be released at https://github.com/OpenNLPLab/AVS-bidirectional

Neural Radiance Fields (NeRF) have demonstrated impressive performance in vision and graphics tasks, such as novel view synthesis and immersive reality. However, the shape-radiance ambiguity of radiance fields remains a challenge, especially in the sparse viewpoints setting. Recent work resorts to integrating depth priors into outdoor NeRF training to alleviate the issue. However, the criteria for selecting depth priors and the relative merits of different priors have not been thoroughly investigated. Moreover, the relative merits of selecting different approaches to use the depth priors is also an unexplored problem. In this paper, we provide a comprehensive study and evaluation of employing depth priors to outdoor neural radiance fields, covering common depth sensing technologies and most application ways. Specifically, we conduct extensive experiments with two representative NeRF methods equipped with four commonly-used depth priors and different depth usages on two widely used outdoor datasets. Our experimental results reveal several interesting findings that can potentially benefit practitioners and researchers in training their NeRF models with depth priors. Project Page: https://cwchenwang.github.io/outdoor-nerf-depth, Comment: Accepted to ACM MM 2023. Project Page: https://cwchenwang.github.io/outdoor-nerf-depth

We propose a latent diffusion model with contrastive learning for audio-visual segmentation (AVS) to extensively explore the contribution of audio. We interpret AVS as a conditional generation task, where audio is defined as the conditional variable for sound producer(s) segmentation. With our new interpretation, it is especially necessary to model the correlation between audio and the final segmentation map to ensure its contribution. We introduce a latent diffusion model to our framework to achieve semantic-correlated representation learning. Specifically, our diffusion model learns the conditional generation process of the ground-truth segmentation map, leading to ground-truth aware inference when we perform the denoising process at the test stage. As a conditional diffusion model, we argue it is essential to ensure that the conditional variable contributes to model output. We then introduce contrastive learning to our framework to learn audio-visual correspondence, which is proven consistent with maximizing the mutual information between model prediction and the audio data. In this way, our latent diffusion model via contrastive learning explicitly maximizes the contribution of audio for AVS. Experimental results on the benchmark dataset verify the effectiveness of our solution. Code and results are online via our project page: https://github.com/OpenNLPLab/DiffusionAVS.

We analysis performance of semantic segmentation models wrt. adversarial attacks, and observe that the adversarial examples generated from a source model fail to attack the target models. i.e The conventional attack methods, such as PGD and FGSM, do not transfer well to target models, making it necessary to study the transferable attacks, especially transferable attacks for semantic segmentation. We find two main factors to achieve transferable attack. Firstly, the attack should come with effective data augmentation and translation-invariant features to deal with unseen models. Secondly, stabilized optimization strategies are needed to find the optimal attack direction. Based on the above observations, we propose an ensemble attack for semantic segmentation to achieve more effective attacks with higher transferability. The source code and experimental results are publicly available via our project page: https://github.com/anucvers/TASS., Comment: Source code is available at: https://github.com/anucvers/TASS

Due to the domain differences and unbalanced disparity distribution across multiple datasets, current stereo matching approaches are commonly limited to a specific dataset and generalize poorly to others. Such domain shift issue is usually addressed by substantial adaptation on costly target-domain ground-truth data, which cannot be easily obtained in practical settings. In this paper, we propose to dig into uncertainty estimation for robust stereo matching. Specifically, to balance the disparity distribution, we employ a pixel-level uncertainty estimation to adaptively adjust the next stage disparity searching space, in this way driving the network progressively prune out the space of unlikely correspondences. Then, to solve the limited ground truth data, an uncertainty-based pseudo-label is proposed to adapt the pre-trained model to the new domain, where pixel-level and area-level uncertainty estimation are proposed to filter out the high-uncertainty pixels of predicted disparity maps and generate sparse while reliable pseudo-labels to align the domain gap. Experimentally, our method shows strong cross-domain, adapt, and joint generalization and obtains \textbf{1st} place on the stereo task of Robust Vision Challenge 2020. Additionally, our uncertainty-based pseudo-labels can be extended to train monocular depth estimation networks in an unsupervised way and even achieves comparable performance with the supervised methods. The code will be available at https://github.com/gallenszl/UCFNet., Comment: Accepted by TPAMI

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.

Relative positional encoding is widely used in vanilla and linear transformers to represent positional information. However, existing encoding methods of a vanilla transformer are not always directly applicable to a linear transformer, because the latter requires a decomposition of the query and key representations into separate kernel functions. Nevertheless, principles for designing encoding methods suitable for linear transformers remain understudied. In this work, we put together a variety of existing linear relative positional encoding approaches under a canonical form and further propose a family of linear relative positional encoding algorithms via unitary transformation. Our formulation leads to a principled framework that can be used to develop new relative positional encoding methods that preserve linear space-time complexity. Equipped with different models, the proposed linearized relative positional encoding (LRPE) family derives effective encoding for various applications. Experiments show that compared with existing methods, LRPE achieves state-of-the-art performance in language modeling, text classification, and image classification. Meanwhile, it emphasizes a general paradigm for designing broadly more relative positional encoding methods that are applicable to linear transformers. The code is available at https://github.com/OpenNLPLab/Lrpe., Comment: Reviewed by TMLR, decision pending. Yiran Zhong is the corresponding author. Code is available at https://github.com/OpenNLPLab/Lrpe

Salient objects attract human attention and usually stand out clearly from their surroundings. In contrast, camouflaged objects share similar colors or textures with the environment. In this case, salient objects are typically non-camouflaged, and camouflaged objects are usually not salient. Due to this inherent contradictory attribute, we introduce an uncertainty-aware learning pipeline to extensively explore the contradictory information of salient object detection (SOD) and camouflaged object detection (COD) via data-level and task-wise contradiction modeling. We first exploit the dataset correlation of these two tasks and claim that the easy samples in the COD dataset can serve as hard samples for SOD to improve the robustness of the SOD model. Based on the assumption that these two models should lead to activation maps highlighting different regions of the same input image, we further introduce a contrastive module with a joint-task contrastive learning framework to explicitly model the contradictory attributes of these two tasks. Different from conventional intra-task contrastive learning for unsupervised representation learning, our contrastive module is designed to model the task-wise correlation, leading to cross-task representation learning. To better understand the two tasks from the perspective of uncertainty, we extensively investigate the uncertainty estimation techniques for modeling the main uncertainties of the two tasks, namely task uncertainty (for SOD) and data uncertainty (for COD), and aiming to effectively estimate the challenging regions for each task to achieve difficulty-aware learning. Experimental results on benchmark datasets demonstrate that our solution leads to both state-of-the-art performance and informative uncertainty estimation.

Unsupervised object discovery (UOD) refers to the task of discriminating the whole region of objects from the background within a scene without relying on labeled datasets, which benefits the task of bounding-box-level localization and pixel-level segmentation. This task is promising due to its ability to discover objects in a generic manner. We roughly categorise existing techniques into two main directions, namely the generative solutions based on image resynthesis, and the clustering methods based on self-supervised models. We have observed that the former heavily relies on the quality of image reconstruction, while the latter shows limitations in effectively modeling semantic correlations. To directly target at object discovery, we focus on the latter approach and propose a novel solution by incorporating weakly-supervised contrastive learning (WCL) to enhance semantic information exploration. We design a semantic-guided self-supervised learning model to extract high-level semantic features from images, which is achieved by fine-tuning the feature encoder of a self-supervised model, namely DINO, via WCL. Subsequently, we introduce Principal Component Analysis (PCA) to localize object regions. The principal projection direction, corresponding to the maximal eigenvalue, serves as an indicator of the object region(s). Extensive experiments on benchmark unsupervised object discovery datasets demonstrate the effectiveness of our proposed solution. The source code and experimental results are publicly available via our project page at https://github.com/npucvr/WSCUOD.git.

In this paper, we present a weakly-supervised RGB-D salient object detection model via scribble supervision. Specifically, as a multimodal learning task, we focus on effective multimodal representation learning via inter-modal mutual information regularization. In particular, following the principle of disentangled representation learning, we introduce a mutual information upper bound with a mutual information minimization regularizer to encourage the disentangled representation of each modality for salient object detection. Based on our multimodal representation learning framework, we introduce an asymmetric feature extractor for our multimodal data, which is proven more effective than the conventional symmetric backbone setting. We also introduce multimodal variational auto-encoder as stochastic prediction refinement techniques, which takes pseudo labels from the first training stage as supervision and generates refined prediction. Experimental results on benchmark RGB-D salient object detection datasets verify both effectiveness of our explicit multimodal disentangled representation learning method and the stochastic prediction refinement strategy, achieving comparable performance with the state-of-the-art fully supervised models. Our code and data are available at: https://github.com/baneitixiaomai/MIRV., Comment: IEEE Transactions on Circuits and Systems for Video Technology 2023

Sequence modeling has important applications in natural language processing and computer vision. Recently, the transformer-based models have shown strong performance on various sequence modeling tasks, which rely on attention to capture pairwise token relations, and position embedding to inject positional information. While showing good performance, the transformer models are inefficient to scale to long input sequences, mainly due to the quadratic space-time complexity of attention. To overcome this inefficiency, we propose to model sequences with a relative position encoded Toeplitz matrix and use a Toeplitz matrix-vector production trick to reduce the space-time complexity of the sequence modeling to log linear. A lightweight sub-network called relative position encoder is proposed to generate relative position coefficients with a fixed budget of parameters, enabling the proposed Toeplitz neural network to deal with varying sequence lengths. In addition, despite being trained on 512-token sequences, our model can extrapolate input sequence length up to 14K tokens in inference with consistent performance. Extensive experiments on autoregressive and bidirectional language modeling, image modeling, and the challenging Long-Range Arena benchmark show that our method achieves better performance than its competitors in most downstream tasks while being significantly faster. The code is available at https://github.com/OpenNLPLab/Tnn., Comment: Accepted to ICLR 2023 Spotlight. Yiran Zhong is the corresponding author. 15B pretrained LLM with TNN will be released at https://github.com/OpenNLPLab/Tnn soon

The normalized eight-point algorithm has been widely viewed as the cornerstone in two-view geometry computation, where the seminal Hartley's normalization has greatly improved the performance of the direct linear transformation algorithm. A natural question is, whether there exists and how to find other normalization methods that may further improve the performance as per each input sample. In this paper, we provide a novel perspective and propose two contributions to this fundamental problem: 1) we revisit the normalized eight-point algorithm and make a theoretical contribution by presenting the existence of different and better normalization algorithms; 2) we introduce a deep convolutional neural network with a self-supervised learning strategy for normalization. Given eight pairs of correspondences, our network directly predicts the normalization matrices, thus learning to normalize each input sample. Our learning-based normalization module can be integrated with both traditional (e.g., RANSAC) and deep learning frameworks (affording good interpretability) with minimal effort. Extensive experiments on both synthetic and real images demonstrate the effectiveness of our proposed approach., Comment: Accepted by Visual Intelligence

This paper discusses the results for the second edition of the Monocular Depth Estimation Challenge (MDEC). This edition was open to methods using any form of supervision, including fully-supervised, self-supervised, multi-task or proxy depth. The challenge was based around the SYNS-Patches dataset, which features a wide diversity of environments with high-quality dense ground-truth. This includes complex natural environments, e.g. forests or fields, which are greatly underrepresented in current benchmarks. The challenge received eight unique submissions that outperformed the provided SotA baseline on any of the pointcloud- or image-based metrics. The top supervised submission improved relative F-Score by 27.62%, while the top self-supervised improved it by 16.61%. Supervised submissions generally leveraged large collections of datasets to improve data diversity. Self-supervised submissions instead updated the network architecture and pretrained backbones. These results represent a significant progress in the field, while highlighting avenues for future research, such as reducing interpolation artifacts at depth boundaries, improving self-supervised indoor performance and overall natural image accuracy., Comment: Published at CVPRW2023

We explore a new task for audio-visual-language modeling called fine-grained audible video description (FAVD). It aims to provide detailed textual descriptions for the given audible videos, including the appearance and spatial locations of each object, the actions of moving objects, and the sounds in videos. Existing visual-language modeling tasks often concentrate on visual cues in videos while undervaluing the language and audio modalities. On the other hand, FAVD requires not only audio-visual-language modeling skills but also paragraph-level language generation abilities. We construct the first fine-grained audible video description benchmark (FAVDBench) to facilitate this research. For each video clip, we first provide a one-sentence summary of the video, ie, the caption, followed by 4-6 sentences describing the visual details and 1-2 audio-related descriptions at the end. The descriptions are provided in both English and Chinese. We create two new metrics for this task: an EntityScore to gauge the completeness of entities in the visual descriptions, and an AudioScore to assess the audio descriptions. As a preliminary approach to this task, we propose an audio-visual-language transformer that extends existing video captioning model with an additional audio branch. We combine the masked language modeling and auto-regressive language modeling losses to optimize our model so that it can produce paragraph-level descriptions. We illustrate the efficiency of our model in audio-visual-language modeling by evaluating it against the proposed benchmark using both conventional captioning metrics and our proposed metrics. We further put our benchmark to the test in video generation models, demonstrating that employing fine-grained video descriptions can create more intricate videos than using captions., Comment: accepted to CVPR 2023, Xuyang Shen, Dong Li and Jinxing Zhou contribute equally, code link: github.com/OpenNLPLab/FAVDBench, dataset link: www.avlbench.opennlplab.cn

Contemporary deep learning multi-scale deblurring models suffer from many issues: 1) They perform poorly on non-uniformly blurred images/videos; 2) Simply increasing the model depth with finer-scale levels cannot improve deblurring; 3) Individual RGB frames contain a limited motion information for deblurring; 4) Previous models have a limited robustness to spatial transformations and noise. Below, we extend the DMPHN model by several mechanisms to address the above issues: I) We present a novel self-supervised event-guided deep hierarchical Multi-patch Network (MPN) to deal with blurry images and videos via fine-to-coarse hierarchical localized representations; II) We propose a novel stacked pipeline, StackMPN, to improve the deblurring performance under the increased network depth; III) We propose an event-guided architecture to exploit motion cues contained in videos to tackle complex blur in videos; IV) We propose a novel self-supervised step to expose the model to random transformations (rotations, scale changes), and make it robust to Gaussian noises. Our MPN achieves the state of the art on the GoPro and VideoDeblur datasets with a 40x faster runtime compared to current multi-scale methods. With 30ms to process an image at 1280x720 resolution, it is the first real-time deep motion deblurring model for 720p images at 30fps. For StackMPN, we obtain significant improvements over 1.2dB on the GoPro dataset by increasing the network depth. Utilizing the event information and self-supervision further boost results to 33.83dB., Comment: International Journal of Computer Vision. arXiv admin note: substantial text overlap with arXiv:1904.03468

Event cameras such as DAVIS can simultaneously output high temporal resolution events and low frame-rate intensity images, which own great potential in capturing scene motion, such as optical flow estimation. Most of the existing optical flow estimation methods are based on two consecutive image frames and can only estimate discrete flow at a fixed time interval. Previous work has shown that continuous flow estimation can be achieved by changing the quantities or time intervals of events. However, they are difficult to estimate reliable dense flow , especially in the regions without any triggered events. In this paper, we propose a novel deep learning-based dense and continuous optical flow estimation framework from a single image with event streams, which facilitates the accurate perception of high-speed motion. Specifically, we first propose an event-image fusion and correlation module to effectively exploit the internal motion from two different modalities of data. Then we propose an iterative update network structure with bidirectional training for optical flow prediction. Therefore, our model can estimate reliable dense flow as two-frame-based methods, as well as estimate temporal continuous flow as event-based methods. Extensive experimental results on both synthetic and real captured datasets demonstrate that our model outperforms existing event-based state-of-the-art methods and our designed baselines for accurate dense and continuous optical flow estimation., Comment: Project page (https://npucvr.github.io/DCEIFlow/). This work has been accepted by IEEE TIP (https://ieeexplore.ieee.org/document/9950520). 15 pages, 10 figures

Although 3D point cloud data has received widespread attentions as a general form of 3D signal expression, applying point clouds to the task of dense correspondence estimation between 3D shapes has not been investigated widely. Furthermore, even in the few existing 3D point cloud-based methods, an important and widely acknowledged principle, i.e . one-to-one matching, is usually ignored. In response, this paper presents a novel end-to-end learning-based method to estimate the dense correspondence of 3D point clouds, in which the problem of point matching is formulated as a zero-one assignment problem to achieve a permutation matching matrix to implement the one-to-one principle fundamentally. Note that the classical solutions of this assignment problem are always non-differentiable, which is fatal for deep learning frameworks. Thus we design a special matching module, which solves a doubly stochastic matrix at first and then projects this obtained approximate solution to the desired permutation matrix. Moreover, to guarantee end-to-end learning and the accuracy of the calculated loss, we calculate the loss from the learned permutation matrix but propagate the gradient to the doubly stochastic matrix directly which bypasses the permutation matrix during the backward propagation. Our method can be applied to both non-rigid and rigid 3D point cloud data and extensive experiments show that our method achieves state-of-the-art performance for dense correspondence learning., Comment: Accepted to IEEE Signal Processing Letters (SPL) 2022

LiDAR depth-only completion is a challenging task to estimate dense depth maps only from sparse measurement points obtained by LiDAR. Even though the depth-only methods have been widely developed, there is still a significant performance gap with the RGB-guided methods that utilize extra color images. We find that existing depth-only methods can obtain satisfactory results in the areas where the measurement points are almost accurate and evenly distributed (denoted as normal areas), while the performance is limited in the areas where the foreground and background points are overlapped due to occlusion (denoted as overlap areas) and the areas where there are no measurement points around (denoted as blank areas) since the methods have no reliable input information in these areas. Building upon these observations, we propose an effective Coupled U-Net (CU-Net) architecture for depth-only completion. Instead of directly using a large network for regression, we employ the local U-Net to estimate accurate values in the normal areas and provide the global U-Net with reliable initial values in the overlap and blank areas. The depth maps predicted by the two coupled U-Nets are fused by learned confidence maps to obtain final results. In addition, we propose a confidence-based outlier removal module, which removes outliers using simple judgment conditions. Our proposed method boosts the final results with fewer parameters and achieves state-of-the-art results on the KITTI benchmark. Moreover, it owns a powerful generalization ability under various depth densities, varying lighting, and weather conditions., Comment: Accepted to IEEE Robotics and Automation Letters (RA-L), Code: https://github.com/YufeiWang777/CU-Net

Existing learning-based point feature descriptors are usually task-agnostic, which pursue describing the individual 3D point clouds as accurate as possible. However, the matching task aims at describing the corresponding points consistently across different 3D point clouds. Therefore these too accurate features may play a counterproductive role due to the inconsistent point feature representations of correspondences caused by the unpredictable noise, partiality, deformation, \etc, in the local geometry. In this paper, we propose to learn a robust task-specific feature descriptor to consistently describe the correct point correspondence under interference. Born with an Encoder and a Dynamic Fusion module, our method EDFNet develops from two aspects. First, we augment the matchability of correspondences by utilizing their repetitive local structure. To this end, a special encoder is designed to exploit two input point clouds jointly for each point descriptor. It not only captures the local geometry of each point in the current point cloud by convolution, but also exploits the repetitive structure from paired point cloud by Transformer. Second, we propose a dynamical fusion module to jointly use different scale features. There is an inevitable struggle between robustness and discriminativeness of the single scale feature. Specifically, the small scale feature is robust since little interference exists in this small receptive field. But it is not sufficiently discriminative as there are many repetitive local structures within a point cloud. Thus the resultant descriptors will lead to many incorrect matches. In contrast, the large scale feature is more discriminative by integrating more neighborhood information. ..., Comment: Accepted to IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) 2022

Vision Transformers have achieved impressive performance in video classification, while suffering from the quadratic complexity caused by the Softmax attention mechanism. Some studies alleviate the computational costs by reducing the number of tokens in attention calculation, but the complexity is still quadratic. Another promising way is to replace Softmax attention with linear attention, which owns linear complexity but presents a clear performance drop. We find that such a drop in linear attention results from the lack of attention concentration on critical features. Therefore, we propose a feature fixation module to reweight the feature importance of the query and key before computing linear attention. Specifically, we regard the query, key, and value as various latent representations of the input token, and learn the feature fixation ratio by aggregating Query-Key-Value information. This is beneficial for measuring the feature importance comprehensively. Furthermore, we enhance the feature fixation by neighborhood association, which leverages additional guidance from spatial and temporal neighbouring tokens. The proposed method significantly improves the linear attention baseline and achieves state-of-the-art performance among linear video Transformers on three popular video classification benchmarks. With fewer parameters and higher efficiency, our performance is even comparable to some Softmax-based quadratic Transformers.

Single image blind deblurring is highly ill-posed as neither the latent sharp image nor the blur kernel is known. Even though considerable progress has been made, several major difficulties remain for blind deblurring, including the trade-off between high-performance deblurring and real-time processing. Besides, we observe that current single image blind deblurring networks cannot further improve or stabilize the performance but significantly degrades the performance when re-deblurring is repeatedly applied. This implies the limitation of these networks in modeling an ideal deblurring process. In this work, we make two contributions to tackle the above difficulties: (1) We introduce the idempotent constraint into the deblurring framework and present a deep idempotent network to achieve improved blind non-uniform deblurring performance with stable re-deblurring. (2) We propose a simple yet efficient deblurring network with lightweight encoder-decoder units and a recurrent structure that can deblur images in a progressive residual fashion. Extensive experiments on synthetic and realistic datasets prove the superiority of our proposed framework. Remarkably, our proposed network is nearly 6.5X smaller and 6.4X faster than the state-of-the-art while achieving comparable high performance., Comment: The first two authors contributed equally, accepted by IEEE TCSVT(https://ieeexplore.ieee.org/document/9868799), Project page(https://npucvr.github.io/IdemDeblur/)

A single rolling-shutter (RS) image may be viewed as a row-wise combination of a sequence of global-shutter (GS) images captured by a (virtual) moving GS camera within the exposure duration. Although RS cameras are widely used, the RS effect causes obvious image distortion especially in the presence of fast camera motion, hindering downstream computer vision tasks. In this paper, we propose to invert the RS image capture mechanism, i.e., recovering a continuous high framerate GS video from two time-consecutive RS frames. We call this task the RS temporal super-resolution (RSSR) problem. The RSSR is a very challenging task, and to our knowledge, no practical solution exists to date. This paper presents a novel deep-learning based solution. By leveraging the multi-view geometry relationship of the RS imaging process, our learning-based framework successfully achieves high framerate GS generation. Specifically, three novel contributions can be identified: (i) novel formulations for bidirectional RS undistortion flows under constant velocity as well as constant acceleration motion model. (ii) a simple linear scaling operation, which bridges the RS undistortion flow and regular optical flow. (iii) a new mutual conversion scheme between varying RS undistortion flows that correspond to different scanlines. Our method also exploits the underlying spatial-temporal geometric relationships within a deep learning framework, where no additional supervision is required beyond the necessary middle-scanline GS image. Building upon these contributions, we represent the very first rolling-shutter temporal super-resolution deep-network that is able to recover high framerate GS videos from just two RS frames. Extensive experimental results on both synthetic and real data show that our proposed method can produce high-quality GS image sequences with rich details, outperforming the state-of-the-art methods., Comment: Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE TPAMI), 16 Pages, 14 Figures

Accurate moving object segmentation is an essential task for autonomous driving. It can provide effective information for many downstream tasks, such as collision avoidance, path planning, and static map construction. How to effectively exploit the spatial-temporal information is a critical question for 3D LiDAR moving object segmentation (LiDAR-MOS). In this work, we propose a novel deep neural network exploiting both spatial-temporal information and different representation modalities of LiDAR scans to improve LiDAR-MOS performance. Specifically, we first use a range image-based dual-branch structure to separately deal with spatial and temporal information that can be obtained from sequential LiDAR scans, and later combine them using motion-guided attention modules. We also use a point refinement module via 3D sparse convolution to fuse the information from both LiDAR range image and point cloud representations and reduce the artifacts on the borders of the objects. We verify the effectiveness of our proposed approach on the LiDAR-MOS benchmark of SemanticKITTI. Our method outperforms the state-of-the-art methods significantly in terms of LiDAR-MOS IoU. Benefiting from the devised coarse-to-fine architecture, our method operates online at sensor frame rate. The implementation of our method is available as open source at: https://github.com/haomo-ai/MotionSeg3D., Comment: Accepted by IROS2022. Code: https://github.com/haomo-ai/MotionSeg3D

Recently, Neural Radiance Fields (NeRF) is revolutionizing the task of novel view synthesis (NVS) for its superior performance. In this paper, we propose to synthesize dynamic scenes. Extending the methods for static scenes to dynamic scenes is not straightforward as both the scene geometry and appearance change over time, especially under monocular setup. Also, the existing dynamic NeRF methods generally require a lengthy per-scene training procedure, where multi-layer perceptrons (MLP) are fitted to model both motions and radiance. In this paper, built on top of the recent advances in voxel-grid optimization, we propose a fast deformable radiance field method to handle dynamic scenes. Our method consists of two modules. The first module adopts a deformation grid to store 3D dynamic features, and a light-weight MLP for decoding the deformation that maps a 3D point in the observation space to the canonical space using the interpolated features. The second module contains a density and a color grid to model the geometry and density of the scene. The occlusion is explicitly modeled to further improve the rendering quality. Experimental results show that our method achieves comparable performance to D-NeRF using only 20 minutes for training, which is more than 70x faster than D-NeRF, clearly demonstrating the efficiency of our proposed method., Comment: Technical Report: 29 pages; project page: https://npucvr.github.io/NDVG

With the ubiquity of rolling shutter (RS) cameras, it is becoming increasingly attractive to recover the latent global shutter (GS) video from two consecutive RS frames, which also places a higher demand on realism. Existing solutions, using deep neural networks or optimization, achieve promising performance. However, these methods generate intermediate GS frames through image warping based on the RS model, which inevitably result in black holes and noticeable motion artifacts. In this paper, we alleviate these issues by proposing a context-aware GS video reconstruction architecture. It facilitates the advantages such as occlusion reasoning, motion compensation, and temporal abstraction. Specifically, we first estimate the bilateral motion field so that the pixels of the two RS frames are warped to a common GS frame accordingly. Then, a refinement scheme is proposed to guide the GS frame synthesis along with bilateral occlusion masks to produce high-fidelity GS video frames at arbitrary times. Furthermore, we derive an approximated bilateral motion field model, which can serve as an alternative to provide a simple but effective GS frame initialization for related tasks. Experiments on synthetic and real data show that our approach achieves superior performance over state-of-the-art methods in terms of objective metrics and subjective visual quality. Code is available at \url{https://github.com/GitCVfb/CVR}., Comment: Accepted to IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2022)

Preys in the wild evolve to be camouflaged to avoid being recognized by predators. In this way, camouflage acts as a key defence mechanism across species that is critical to survival. To detect and segment the whole scope of a camouflaged object, camouflaged object detection (COD) is introduced as a binary segmentation task, with the binary ground truth camouflage map indicating the exact regions of the camouflaged objects. In this paper, we revisit this task and argue that the binary segmentation setting fails to fully understand the concept of camouflage. We find that explicitly modeling the conspicuousness of camouflaged objects against their particular backgrounds can not only lead to a better understanding about camouflage, but also provide guidance to designing more sophisticated camouflage techniques. Furthermore, we observe that it is some specific parts of camouflaged objects that make them detectable by predators. With the above understanding about camouflaged objects, we present the first triple-task learning framework to simultaneously localize, segment, and rank camouflaged objects, indicating the conspicuousness level of camouflage. As no corresponding datasets exist for either the localization model or the ranking model, we generate localization maps with an eye tracker, which are then processed according to the instance level labels to generate our ranking-based training and testing dataset. We also contribute the largest COD testing set to comprehensively analyse performance of the COD models. Experimental results show that our triple-task learning framework achieves new state-of-the-art, leading to a more explainable COD network. Our code, data, and results are available at: \url{https://github.com/JingZhang617/COD-Rank-Localize-and-Segment}., Comment: IEEE Transactions on Circuits and Systems for Video Technology 2023

Directly regressing the non-rigid shape and camera pose from the individual 2D frame is ill-suited to the Non-Rigid Structure-from-Motion (NRSfM) problem. This frame-by-frame 3D reconstruction pipeline overlooks the inherent spatial-temporal nature of NRSfM, i.e., reconstructing the whole 3D sequence from the input 2D sequence. In this paper, we propose to model deep NRSfM from a sequence-to-sequence translation perspective, where the input 2D frame sequence is taken as a whole to reconstruct the deforming 3D non-rigid shape sequence. First, we apply a shape-motion predictor to estimate the initial non-rigid shape and camera motion from a single frame. Then we propose a context modeling module to model camera motions and complex non-rigid shapes. To tackle the difficulty in enforcing the global structure constraint within the deep framework, we propose to impose the union-of-subspace structure by replacing the self-expressiveness layer with multi-head attention and delayed regularizers, which enables end-to-end batch-wise training. Experimental results across different datasets such as Human3.6M, CMU Mocap and InterHand prove the superiority of our framework., Comment: has been accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence

3D point cloud registration is fragile to outliers, which are labeled as the points without corresponding points. To handle this problem, a widely adopted strategy is to estimate the relative pose based only on some accurate correspondences, which is achieved by building correspondences on the identified inliers or by selecting reliable ones. However, these approaches are usually complicated and time-consuming. By contrast, the virtual point-based methods learn the virtual corresponding points (VCPs) for all source points uniformly without distinguishing the outliers and the inliers. Although this strategy is time-efficient, the learned VCPs usually exhibit serious collapse degeneration due to insufficient supervision and the inherent distribution limitation. In this paper, we propose to exploit the best of both worlds and present a novel robust 3D point cloud registration framework. We follow the idea of the virtual point-based methods but learn a new type of virtual points called rectified virtual corresponding points (RCPs), which are defined as the point set with the same shape as the source and with the same pose as the target. Hence, a pair of consistent point clouds, i.e. source and RCPs, is formed by rectifying VCPs to RCPs (VRNet), through which reliable correspondences between source and RCPs can be accurately obtained. Since the relative pose between source and RCPs is the same as the relative pose between source and target, the input point clouds can be registered naturally. Specifically, we first construct the initial VCPs by using an estimated soft matching matrix to perform a weighted average on the target points. Then, we design a correction-walk module to learn an offset to rectify VCPs to RCPs, which effectively breaks the distribution limitation of VCPs. Finally, we develop a hybrid loss function to enforce the shape and geometry structure consistency ..., Comment: Accepted by IEEE Transactions on Circuits and Systems for Video Technology

3D point cloud registration in remote sensing field has been greatly advanced by deep learning based methods, where the rigid transformation is either directly regressed from the two point clouds (correspondences-free approaches) or computed from the learned correspondences (correspondences-based approaches). Existing correspondences-free methods generally learn the holistic representation of the entire point cloud, which is fragile for partial and noisy point clouds. In this paper, we propose a correspondences-free unsupervised point cloud registration (UPCR) method from the representation separation perspective. First, we model the input point cloud as a combination of pose-invariant representation and pose-related representation. Second, the pose-related representation is used to learn the relative pose wrt a "latent canonical shape" for the source and target point clouds respectively. Third, the rigid transformation is obtained from the above two learned relative poses. Our method not only filters out the disturbance in pose-invariant representation but also is robust to partial-to-partial point clouds or noise. Experiments on benchmark datasets demonstrate that our unsupervised method achieves comparable if not better performance than state-of-the-art supervised registration methods., Comment: Accepted by IEEE Geoscience and Remote Sensing Letters

The task of semi-supervised video object segmentation (VOS) has been greatly advanced and state-of-the-art performance has been made by dense matching-based methods. The recent methods leverage space-time memory (STM) networks and learn to retrieve relevant information from all available sources, where the past frames with object masks form an external memory and the current frame as the query is segmented using the mask information in the memory. However, when forming the memory and performing matching, these methods only exploit the appearance information while ignoring the motion information. In this paper, we advocate the return of the \emph{motion information} and propose a motion uncertainty-aware framework (MUNet) for semi-supervised VOS. First, we propose an implicit method to learn the spatial correspondences between neighboring frames, building upon a correlation cost volume. To handle the challenging cases of occlusion and textureless regions during constructing dense correspondences, we incorporate the uncertainty in dense matching and achieve motion uncertainty-aware feature representation. Second, we introduce a motion-aware spatial attention module to effectively fuse the motion feature with the semantic feature. Comprehensive experiments on challenging benchmarks show that \textbf{\textit{using a small amount of data and combining it with powerful motion information can bring a significant performance boost}}. We achieve ${76.5\%}$ $\mathcal{J} \& \mathcal{F}$ only using DAVIS17 for training, which significantly outperforms the \textit{SOTA} methods under the low-data protocol. \textit{The code will be released.}

Salient object detection is subjective in nature, which implies that multiple estimations should be related to the same input image. Most existing salient object detection models are deterministic following a point to point estimation learning pipeline, making them incapable of estimating the predictive distribution. Although latent variable model based stochastic prediction networks exist to model the prediction variants, the latent space based on the single clean saliency annotation is less reliable in exploring the subjective nature of saliency, leading to less effective saliency divergence modeling. Given multiple saliency annotations, we introduce a general divergence modeling strategy via random sampling, and apply our strategy to an ensemble based framework and three latent variable model based solutions to explore the subjective nature of saliency. Experimental results prove the superior performance of our general divergence modeling strategy., Comment: Code is available at: https://npucvr.github.io/Divergence_SOD/

Uncertainty estimation has been extensively studied in recent literature, which can usually be classified as aleatoric uncertainty and epistemic uncertainty. In current aleatoric uncertainty estimation frameworks, it is often neglected that the aleatoric uncertainty is an inherent attribute of the data and can only be correctly estimated with an unbiased oracle model. Since the oracle model is inaccessible in most cases, we propose a new sampling and selection strategy at train time to approximate the oracle model for aleatoric uncertainty estimation. Further, we show a trivial solution in the dual-head based heteroscedastic aleatoric uncertainty estimation framework and introduce a new uncertainty consistency loss to avoid it. For epistemic uncertainty estimation, we argue that the internal variable in a conditional latent variable model is another source of epistemic uncertainty to model the predictive distribution and explore the limited knowledge about the hidden true model. We validate our observation on a dense prediction task, i.e., camouflaged object detection. Our results show that our solution achieves both accurate deterministic results and reliable uncertainty estimation.

Even though considerable progress has been made in deep learning-based 3D point cloud processing, how to obtain accurate correspondences for robust registration remains a major challenge because existing hard assignment methods cannot deal with outliers naturally. Alternatively, the soft matching-based methods have been proposed to learn the matching probability rather than hard assignment. However, in this paper, we prove that these methods have an inherent ambiguity causing many deceptive correspondences. To address the above challenges, we propose to learn a partial permutation matching matrix, which does not assign corresponding points to outliers, and implements hard assignment to prevent ambiguity. However, this proposal poses two new problems, i.e., existing hard assignment algorithms can only solve a full rank permutation matrix rather than a partial permutation matrix, and this desired matrix is defined in the discrete space, which is non-differentiable. In response, we design a dedicated soft-to-hard (S2H) matching procedure within the registration pipeline consisting of two steps: solving the soft matching matrix (S-step) and projecting this soft matrix to the partial permutation matrix (H-step). Specifically, we augment the profit matrix before the hard assignment to solve an augmented permutation matrix, which is cropped to achieve the final partial permutation matrix. Moreover, to guarantee end-to-end learning, we supervise the learned partial permutation matrix but propagate the gradient to the soft matrix instead. Our S2H matching procedure can be easily integrated with existing registration frameworks, which has been verified in representative frameworks including DCP, RPMNet, and DGR. Extensive experiments have validated our method, which creates a new state-of-the-art performance for robust 3D point cloud registration. The code will be made public., Comment: Accepted by AAAI 2022

Deep neural networks can be roughly divided into deterministic neural networks and stochastic neural networks.The former is usually trained to achieve a mapping from input space to output space via maximum likelihood estimation for the weights, which leads to deterministic predictions during testing. In this way, a specific weights set is estimated while ignoring any uncertainty that may occur in the proper weight space. The latter introduces randomness into the framework, either by assuming a prior distribution over model parameters (i.e. Bayesian Neural Networks) or including latent variables (i.e. generative models) to explore the contribution of latent variables for model predictions, leading to stochastic predictions during testing. Different from the former that achieves point estimation, the latter aims to estimate the prediction distribution, making it possible to estimate uncertainty, representing model ignorance about its predictions. We claim that conventional deterministic neural network based dense prediction tasks are prone to overfitting, leading to over-confident predictions, which is undesirable for decision making. In this paper, we investigate stochastic neural networks and uncertainty estimation techniques to achieve both accurate deterministic prediction and reliable uncertainty estimation. Specifically, we work on two types of uncertainty estimations solutions, namely ensemble based methods and generative model based methods, and explain their pros and cons while using them in fully/semi/weakly-supervised framework. Due to the close connection between uncertainty estimation and model calibration, we also introduce how uncertainty estimation can be used for deep model calibration to achieve well-calibrated models, namely dense model calibration. Code and data are available at https://github.com/JingZhang617/UncertaintyEstimation., Comment: Technical Report

Existing RGB-D saliency detection models do not explicitly encourage RGB and depth to achieve effective multi-modal learning. In this paper, we introduce a novel multi-stage cascaded learning framework via mutual information minimization to "explicitly" model the multi-modal information between RGB image and depth data. Specifically, we first map the feature of each mode to a lower dimensional feature vector, and adopt mutual information minimization as a regularizer to reduce the redundancy between appearance features from RGB and geometric features from depth. We then perform multi-stage cascaded learning to impose the mutual information minimization constraint at every stage of the network. Extensive experiments on benchmark RGB-D saliency datasets illustrate the effectiveness of our framework. Further, to prosper the development of this field, we contribute the largest (7x larger than NJU2K) dataset, which contains 15,625 image pairs with high quality polygon-/scribble-/object-/instance-/rank-level annotations. Based on these rich labels, we additionally construct four new benchmarks with strong baselines and observe some interesting phenomena, which can motivate future model design. Source code and dataset are available at "https://github.com/JingZhang617/cascaded_rgbd_sod"., Comment: Accepted as ICCV2021 paper

We propose PR-RRN, a novel neural-network based method for Non-rigid Structure-from-Motion (NRSfM). PR-RRN consists of Residual-Recursive Networks (RRN) and two extra regularization losses. RRN is designed to effectively recover 3D shape and camera from 2D keypoints with novel residual-recursive structure. As NRSfM is a highly under-constrained problem, we propose two new pairwise regularization to further regularize the reconstruction. The Rigidity-based Pairwise Contrastive Loss regularizes the shape representation by encouraging higher similarity between the representations of high-rigidity pairs of frames than low-rigidity pairs. We propose minimum singular-value ratio to measure the pairwise rigidity. The Pairwise Consistency Loss enforces the reconstruction to be consistent when the estimated shapes and cameras are exchanged between pairs. Our approach achieves state-of-the-art performance on CMU MOCAP and PASCAL3D+ dataset., Comment: Accepted to ICCV 2021

The vast majority of modern consumer-grade cameras employ a rolling shutter mechanism, leading to image distortions if the camera moves during image acquisition. In this paper, we present a novel deep network to solve the generic rolling shutter correction problem with two consecutive frames. Our pipeline is symmetrically designed to predict the global shutter image corresponding to the intermediate time of these two frames, which is difficult for existing methods because it corresponds to a camera pose that differs most from the two frames. First, two time-symmetric dense undistortion flows are estimated by using well-established principles: pyramidal construction, warping, and cost volume processing. Then, both rolling shutter images are warped into a common global shutter one in the feature space, respectively. Finally, a symmetric consistency constraint is constructed in the image decoder to effectively aggregate the contextual cues of two rolling shutter images, thereby recovering the high-quality global shutter image. Extensive experiments with both synthetic and real data from public benchmarks demonstrate the superiority of our proposed approach over the state-of-the-art methods., Comment: Accepted by IEEE International Conference on Computer Vision (ICCV) 2021