Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Mingyi HeSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works239 from public data
- 1,955
- Deep Stacked Hierarchical Multi-Patch Network for Image Deblurring663
A deep {hierarchical multi-patch network} inspired by Spatial Pyramid Matching to deal with blurry images via a fine-to-coarse hierarchical representation and is the first real-time deep motion deblurring model for 720p images at 30fps.
- Rotation Averaging650
This paper is conceived as a tutorial on rotation averaging, summarizing the research that has been carried out in this area and contains many new results, which were developed to fill gaps in knowledge, answering fundamental questions such as radius of convergence of the algorithms, and existence of local minima.
- 647
- Simultaneously Localize, Segment and Rank the Camouflaged Objects599
It is argued that explicitly modeling the conspicuousness of camouflaged objects against their particular backgrounds can not only lead to a better understanding about camouflage and evolution of animals, but also provide guidance to design more sophisticated camouflage techniques.
- IoU Loss for 2D/3D Object Detection492
By integrating the implemented IoU loss into several state-of-the-art 3D object detectors, consistent improvements have been achieved for both bird-eye-view 2D detection and point cloud 3D detection on the public KITTI [3] benchmark.
- Hierarchical Neural Architecture Search for Deep Stereo Matching438
This paper proposes the first end-to-end hierarchical NAS framework for deep stereo matching by incorporating task-specific human knowledge into the neural architecture search framework and optimizing the architectures of the entire pipeline jointly.
- CFNet: Cascade and Fused Cost Volume for Robust Stereo Matching406
This paper proposes CFNet, a Cascade and Fused cost volume based network to improve the robustness of the stereo matching network, and employs a variance-based uncertainty estimation to adaptively adjust the next stage disparity search space.
- UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders396
Inspired by the saliency data labeling process, a probabilistic RGB-D saliency detection network via conditional variational autoencoders to model human annotation uncertainty and generate multiple saliency maps for each input image by sampling in the latent space is proposed.
- 373
- Bringing a Blurry Frame Alive at High Frame-Rate With an Event Camera341
This paper proposes a simple and effective approach, the Event-based Double Integral (EDI) model, to reconstruct a high frame-rate, sharp video from a single blurry frame and its event data, based on solving a simple non-convex optimization problem in a single scalar variable.
- Weakly-Supervised Salient Object Detection via Scribble Annotations320
This paper proposes a weakly-supervised salient object detection model to learn saliency from scribble annotations, and presents a new metric, termed saliency structure measure, as a complementary metric to evaluate sharpness of the prediction.
- Uncertainty-aware Joint Salient Object and Camouflaged Object Detection308
This paper proposes a paradigm of lever-aging the contradictory information to enhance the detection ability of both salient object detection and camouflaged object detection, and introduces a "similarity measure" module to explicitly model the contradicting attributes of these two tasks.
- Skeleton based action recognition using translation-scale invariant image mapping and multi-scale deep CNN260
An image classification based approach to large scale action recognition from 3D skeleton videos is presented and a multi-scale deep convolutional neural network (CNN) is proposed for the image classification task, which could enhance the temporal frequency adjustment of the model.
- ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving186
This paper contributes the first large scale database suitable for 3D car instance understanding – ApolloCar3D and builds various baseline algorithms with the state-of-the-art deep convolutional neural networks.
- 181
- Complementary Patch for Weakly Supervised Semantic Segmentation173
A novel Complementary Patch (CP) Representation is proposed and it is proved that the information of the sum of the CAMs by a pair of input images with complementary hidden (patched) parts is greater than or equal to theInformation of the baseline CAM.
- Uncertainty Inspired RGB-D Saliency Detection168
Qualitative and quantitative results on six challengingRGB-D benchmark datasets show the first stochastic framework to employ uncertainty for RGB-D saliency detection by learning from the data labeling process superior performance in learning the distribution of saliency maps.
- Deep learning based point cloud registration: an overview157
This paper summarizes recent progress and presents a comprehensive overview of popular deep learning based point cloud registration methods, and classify the popular approaches into different categories such as, correspondences-based or correspondence-free, effective modules: feature extractor, matching, outlier rejection, and motion estimation.
- Monocular depth estimation with hierarchical fusion of dilated CNNs and soft-weighted-sum inference145
A deep end-to-end learning framework to tackle monocular depth estimation, which learns the direct mapping from a color image to the corresponding depth map, and utilizes soft-weighted-sum inference instead of the hard-max inference, transforming the discretized depth score to continuous depth value.
- 3D Geometry-aware Deformable Gaussian Splatting for Dynamic View Synthesis143
This paper proposes a 3D geometry-aware deformable Gaussian Splatting method for dynamic view synthesis that achieves 3D geometry-aware deformation modeling, which enables improved dynamic view synthesis and 3D dynamic reconstruction.
- RGB-D Saliency Detection via Cascaded Mutual Information Minimization131
This paper introduces a novel multistage cascaded learning framework via mutual information minimization to explicitly model the multi-modal information between RGB image and depth data.
- 130
- PCW-Net: Pyramid Combination and Warping Cost Volume for Stereo Matching125
The proposed PCW-Net is a Pyramid Combination and Warping cost volume-based network to achieve good performance on both cross-domain generalization and stereo matching accuracy on various benchmarks, and shows strong cross- domaingeneralization and outperforms existing state-of-the-arts with a large margin.
- 124
- Learning Dense and Continuous Optical Flow From an Event Camera118
This paper proposes 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 and outperforms existing event-based state-of-the-art methods.
- Toward Deeper Understanding of Camouflaged Object Detection115
This paper argues that the binary segmentation setting fails to fully understand the concept of camouflage, and presents the first triple-task learning framework to simultaneously localize, segment, and rank camouflaged objects, indicating the conspicuousness level of camouflage.
- 109
- 109
- LRRU: Long-short Range Recurrent Updating Networks for Depth Completion104
This work proposes a novel lightweight deep network framework, the Long-short Range Recurrent Updating (LRRU) network, which outperforms competing approaches on the NYUv2 dataset, and ranks 1st on the KITTI depth completion benchmark at the time of submission.
- 104
- MVS2: Deep Unsupervised Multi-View Stereo with Multi-View Symmetry97
This paper proposes the first unsupervised learning based MVS network, which learns the multi-view depth maps from the input multi- view images and does not need ground-truth 3D training data.
- Channel Attention Based Iterative Residual Learning for Depth Map Super-Resolution95
This paper proposes a new framework for real-world DSR, which consists of four modules: an iterative residual learning module with deep supervision to learn effective high-frequency components of depth maps in a coarse-to-fine manner, and a depth refinement module to improve the depth map by TGV regularization and input loss.
- Efficient Spatial-Temporal Information Fusion for LiDAR-Based 3D Moving Object Segmentation94
This work proposes a novel deep neural network exploiting both spatial-temporal information and different representation modalities ofLiDAR scans to improve LiDAR-MOS performance and outperforms the state-of-the-art methods significantly in terms of LiDARS IoU.
- Deep Depth Super-Resolution: Learning Depth Super-Resolution Using Deep Convolutional Neural Network91
The proposed deep depth super-resolution method learns the mapping from a low-resolution depth image to a high-resolution one in an end-to-end style and outperforms the state-of-the-art depth image super- resolution methods with a margin.
- Forward Flow for Novel View Synthesis of Dynamic Scenes90
Thorough experiments show that the proposed neural radiance field approach outperforms existing methods in both novel view rendering and motion modeling, demonstrating the effectiveness of the forward flow motion modeling.
- Rolling Shutter Camera Relative Pose: Generalized Epipolar Geometry84
The rolling shutter essential matrix is introduced, and its link to existing models such as the push-broom cameras is established, summarized in a clean hierarchy of multi-perspective cameras.
- Open-World Stereo Video Matching with Deep RNN79
A novel Recurrent Neural Network (RNN) that takes a continuous stereo video as input, and directly predicts a depth-map at each frame without a pre-training process, and without the need of ground-truth depth-maps as supervision.
- Monocular Dense 3D Reconstruction of a Complex Dynamic Scene from Two Perspective Frames76
This paper proposes a new approach for monocular dense 3D reconstruction of a complex dynamic scene from two perspective frames by applying superpixel over-segmentation to the image, and model a generically dynamic scene with a piecewise planar and rigid approximation.
- Noise-Aware Unsupervised Deep Lidar-Stereo Fusion75
LidarStereoNet is presented, the first unsupervised Lidar-stereo fusion network, which can be trained in an end-to-end manner without the need of ground truth depth maps, and is proposed to incorporate the piecewise planar model into the network learning to further constrain depths to conform to the underlying 3D geometry.
- Unsupervised Deep Epipolar Flow for Stationary or Dynamic Scenes75
Experimental results on various benchmarking datasets show that the Deep Epipolar Flow method achieves competitive performance compared with supervised methods and outperforms state-of-the-art unsupervised deep-learning methods.
- 74
- Phase-Only Image Based Kernel Estimation for Single Image Blind Deblurring73
It is shown that the auto-correlation of the absolute phase-only image 1 can provide faithful information about the motion that caused the blur, leading to a new and efficient blur kernel estimation approach.
- Ground-Plane-Based Absolute Scale Estimation for Monocular Visual Odometry70
This paper proposes a robust divide-and-conquer absolute scale estimation method based on the ground plane and camera height by analyzing the advantages and disadvantages of different approaches.
- Deep Two-View Structure-from-Motion Revisited69
This work proposes to revisit the problem of deep two-view SfM by leveraging the well-posedness of the classic pipeline by developing a scale-invariant depth estimation network that leverages epipolar geometry to reduce the search space, refine the dense correspondences, and estimate relative depth maps.
- 68
- VRNet: Learning the Rectified Virtual Corresponding Points for 3D Point Cloud Registration67
This paper proposes a novel robust 3D point cloud registration framework that achieves advanced registration performance and time-efficiency simultaneously and develops a hybrid loss function to enforce the shape and geometry structure consistency of the learned RCPs and the source to provide sufficient supervision.
- Generative Transformer for Accurate and Reliable Salient Object Detection66
This work introduces a latent variable model within the transformer framework, termed the inferential generative adversarial network (iGAN), and applies it to fully supervised salient object detection, explaining that iGAN within the transformer framework leads to both accurate and reliable salient object detection.
- 65
- Simultaneous Stereo Video Deblurring and Scene Flow Estimation65
This paper proposes a novel approach to deblurring from stereo videos that exploits the piece-wise planar assumption about the scene and leverage the scene flow information todeblur the image and achieves significant improvement in flow estimation and removing the blur effect over the state-of theart methods.
- Multimodal Variational Auto-encoder based Audio-Visual Segmentation64
An Explicit Conditional Multimodal Variational Auto-Encoder for audio-visual segmentation (AVS), aiming to segment sound sources in the video sequence, with a new state-of-the-art performance leap on the challenging MS3 subset for multiple sound source segmentation.
- Scalable Dense Non-rigid Structure-from-Motion: A Grassmannian Perspective63
This paper proposes a new approach for dense NRSfM by modeling the problem on a Grassmann manifold that is scalable and more accurate than state-of-the-art methods, and more robust to noise and generalizes to highly nonlinear deformations.
- 63
- NLCA-Net: a non-local context attention network for stereo matching59
A simple yet effective non-local context attention network to exploit the global context information by using attention mechanisms and semantic information for stereo matching is proposed.
- 56
- Spatio-temporal union of subspaces for multi-body non-rigid structure-from-motion56
This paper proposes a unified framework to jointly segment and reconstruct multiple non-rigid objects and proposes to exploit the structure of the scenes along both temporal direction and spatial direction, thus achieving a spatio-temporal representation.
- Mutual Information Regularization for Weakly-Supervised RGB-D Salient Object Detection52
This paper presents a weakly-supervised RGB-D salient object detection model via scribble supervision, and introduces multimodal variational auto-encoder as stochastic prediction refinement techniques, which takes pseudo labels from the first training stage as supervision and generates refined prediction.
- 52
- Deeply Supervised Depth Map Super-Resolution as Novel View Synthesis52
To handle large up-sampling factors, a deeply supervised network structure is presented to enforce strong supervision in each stage of the network and a multi-scale fusion strategy is proposed to effectively exploit the feature maps at different scales and handle the blocking effect.
- Multi-Body Non-Rigid Structure-from-Motion52
This paper presents the first multi-body non-rigid structure-from-motion (SFM) method, which simultaneously reconstructs and segments multiple objects that are undergoing non- Rigid deformation over time.
- Digging Into Uncertainty-Based Pseudo-Label for Robust Stereo Matching51
This paper proposes to dig into uncertainty estimation for robust stereo matching by employing a pixel-level uncertainty estimation to adaptively adjust the next stage disparity searching space and an uncertainty-based pseudo-label is proposed to adapt the pre-trained model to the new domain.
- 47
- 47
- 45
- SUNet: Symmetric Undistortion Network for Rolling Shutter Correction44
This paper presents a novel deep network to solve the generic rolling shutter correction problem with two consecutive frames, symmetrically designed to predict the global shutter image corresponding to the intermediate time of these two frames.
- 43
- 42
- Robust Motion Segmentation with Unknown Correspondences41
This paper introduces an approach to performing motion segmentation without any prior knowledge of point correspondences in terms of Partial Permutation Matrices (PPMs) and aims to match feature descriptors while simultaneously encouraging point trajectories to satisfy subspace constraints.
- Deep Idempotent Network for Efficient Single Image Blind Deblurring40
A deep idempotent network is presented to achieve improved blind non-uniform deblurring performance with stable re-deblurring and a simple yet efficient deblurred network with lightweight encoder-decoder units and a recurrent structure that can deblur images in a progressive residual fashion is proposed.
- Superpixel Soup: Monocular Dense 3D Reconstruction of a Complex Dynamic Scene40
This work proposes a unified approach to solve the task of dense 3D reconstruction of a complex dynamic scene from images and shows that this approach provides an effective solution to the inherent scale-ambiguity in structure-from-motion under perspective projection.
- 3D skeleton based action recognition by video-domain translation-scale invariant mapping and multi-scale dilated CNN39
A video domain translation-scale invariant image mapping is proposed, which transforms the 3D skeleton videos to color images, namely skeleton images, and a multi-scale dilated convolutional neural network is designed for the classification of the skeleton images.
- 38
- 36
- 36
- 33
- 33
- 33
- Element-Wise Factorization for N-View Projective Reconstruction33
This paper recasts the problem as matrix factorization as element-wise factorization, leading to a convenient and efficient semi-definite program formulation, and presents an optimal solution to projective factorization for structure and motion, based on the same principle of low-rank factorization.
- End-to-End Learning the Partial Permutation Matrix for Robust 3D Point Cloud Registration32
A dedicated soft-to-hard (S2H) matching procedure within the registration pipeline consisting of two steps: solving the soft matching matrix and projecting this soft matrix to the partial permutation matrix (H-step), which creates a new state-of-the-art performance.
- Decomposed Guided Dynamic Filters for Efficient RGB-Guided Depth Completion31
This paper proposes to decompose the guided dynamic filters into a spatially-shared component multiplied by content-adaptive adaptors at each spatial location and introduces two decomposition schemes, which decompose the filters by splitting the filter structure and using spatial-wise attention, respectively.
- Learning Saliency From Single Noisy Labelling: A Robust Model Fitting Perspective31
This paper calls on the theory of robust model fitting and formulate deep saliency prediction from a single noisy labelling as robust network learning and exploit model consistency across iterations to identify inliers and outliers.
- 30
- Event-guided Multi-patch Network with Self-supervision for Non-uniform Motion Deblurring30
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 and achieves the state of the art on the GoPro and VideoDeblur datasets with a 40% faster runtime compared to current multi-scale methods.
- 29
- 29
- 29
- 28
- 28
- 27
- 26
- 26
- 26
- Rolling Shutter Inversion: Bring Rolling Shutter Images to High Framerate Global Shutter Video25
This paper represents the very first rolling-shuttered temporal super-resolution deep-network that is able to recover high framerate global-shutter videos from just two RS frames, and exploits the underlying spatial-temporal geometric relationships within a deep learning framework.
- “Maximizing Rigidity” Revisited: A Convex Programming Approach for Generic 3D Shape Reconstruction from Multiple Perspective Views23
This work revisits the principle of “maximizing rigidity” in structure-from-motion literature, and develops a unified theory which is applicable to both rigid and non-rigid structure reconstruction in a rigidity-agnostic way.
- 22
- 21
- 21
- 20
- 19
- 19
- 19
- 18
- Deep Non-Rigid Structure-From-Motion: A Sequence-to-Sequence Translation Perspective17
The Context Layer is proposed, which enables the deep learning framework to effectively impose overall constraints on sequences based on the structural characteristics of non-rigid sequences based on the structural characteristics of non-rigid sequences.
- 17
- 17
- 17
- 17
- 16
- The Second Monocular Depth Estimation Challenge16
The challenge received eight unique submissions that outperformed the provided SotA baseline on any of the pointcloud- or image-based metrics, and the top supervised submission improved relative F-Score, while the top self-supervised improved it by 16.61%.
- 16
- Depth Map Completion by Jointly Exploiting Blurry Color Images and Sparse Depth Maps16
This paper proposes to tackle the problem of depth map completion by jointly exploiting the blurry color image sequences and the sparse depth map measurements, and presents an energy minimization based formulation to simultaneously complete the depth maps, estimate the scene flow and deblur the color images.
- Stereo Computation for a Single Mixture Image16
This work gives a novel deep-learning based solution to the original problem of stereo computation from a single mixture image, by jointly solving the two subtasks of image layer separation as well as stereo matching.
- 15
- 15
- 14
- 14
- 14
- 14
- 14
- 14
- 13
- 13
- 3D Geometry-Aware Semantic Labeling of Outdoor Street Scenes13
This paper proposes a 3D semantic labeling method to label outdoor street scenes whenever a dense depth map is available, and demonstrates that directly performing 3D convolution in the framework of a residual connected 3D voxel top-down modulation network can lead to superior results.
- 12
- 12
- 12
- 11
- Unsupervised 3D Pose Estimation with Non-Rigid Structure-from-Motion Modeling11
A new modeling method for human pose deformations and an accompanying diffusion-based motion prior is proposed and evaluated on mainstream datasets and obtained superior results outperforming the state-of-the-art.
- 11
- 11
- 11
- Integrated deep and shallow networks for salient object detection11
This paper takes results of unsupervised saliency and normalized color images as inputs, and directly learns an end-to-end mapping between inputs and the corresponding saliency maps to obtain a spatially consistent saliency map with sharp object boundaries.
- 11
- 10
- 10
- 9
- Displacement-Invariant Cost Computation for Stereo Matching9
This work proposes a displacement-invariant cost computation module to compute the matching costs without needing a 4D feature volume, and proposes an entropy-based refinement strategy to refine the computed disparity map, which further improves the speed by avoiding the need to compute a second disparity map on the right image.
- 9
- 9
- 9
- 8
- 8
- 8
- 8
- 8
- 8
- 7
- 7
- 7
- 7
- 6
- 6
- 6
- 6
- 6
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- Deep Stochastic Attraction and Repulsion Embedding for Image Based Localization.5
This paper represents a place as a set of exemplar images depicting the same landmarks, instead of some pre-defined geographic locations by partitioning the world, and proposes a new Stochastic Attraction and Repulsion Embedding loss function to facilitate the competitive learning.
- 5
- 5
- 4
- 4
- 4
- 4
- A Motion Free Approach to Dense Depth Estimation in Complex Dynamic Scene.4
This work shows that, given per-pixel optical flow correspondences between two consecutive frames and the sparse depth prior for the reference frame, it can recover the dense depth map for the successive frames without solving for motion parameters.
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-13. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.