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
- Chi–Keung TangSuggested from co-authorship
- Heung‐Yeung ShumSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works314 from public data
- Pyramid Scene Parsing Network15,291
This paper exploits the capability of global context information by different-region-based context aggregation through the pyramid pooling module together with the proposed pyramid scene parsing network (PSPNet) to produce good quality results on the scene parsing task.
- Path Aggregation Network for Instance Segmentation9,307
Path Aggregation Network (PANet) is proposed aiming at boosting information flow in proposal-based instance segmentation framework by enhancing the entire feature hierarchy with accurate localization signals in lower layers by bottom-up path augmentation.
- Point Transformer2,467
The Point Transformer design improves upon prior work across domains and tasks and crosses the 70% mIoU threshold for the first time on the challenging S3DIS dataset.
- 1,797
- ICNet for Real-Time Semantic Segmentation on High-Resolution Images1,627
An image cascade network (ICNet) that incorporates multi-resolution branches under proper label guidance to address the challenging task of real-time semantic segmentation is proposed and in-depth analysis of the framework is provided.
- Image smoothing via L 0 gradient minimization1,508
This work presents a new image editing method, particularly effective for sharpening major edges by increasing the steepness of transition while eliminating a manageable degree of low-amplitude structures in an optimization framework making use of L0 gradient minimization.
- 1,435
- 1,386
- Scale-Recurrent Network for Deep Image Deblurring1,338
A Scale-recurrent Network (SRN-DeblurNet) is proposed and shown to produce better quality results than state-of-the-arts, both quantitatively and qualitatively in single image deblurring.
- 3DSSD: Point-Based 3D Single Stage Object Detector1,220
This paper presents a lightweight point-based 3D single stage object detector 3DSSD to achieve decent balance of accuracy and efficiency, and outperforms all state-of-the-art voxel-based single-stage methods by a large margin.
- PSANet: Point-wise Spatial Attention Network for Scene Parsing1,191
The point-wise spatial attention network (PSANet) is proposed to relax the local neighborhood constraint and achieves top performance on various competitive scene parsing datasets, including ADE20K, PASCAL VOC 2012 and Cityscapes, demonstrating its effectiveness and generality.
- 1,159
- ScribbleSup: Scribble-Supervised Convolutional Networks for Semantic Segmentation1,154
This paper proposes to use scribbles to annotate images, and develops an algorithm to train convolutional networks for semantic segmentation supervised by scribbles, which shows excellent results on the PASCALCONTEXT dataset thanks to extra inexpensive scribble annotations.
- Two-Phase Kernel Estimation for Robust Motion Deblurring1,147
It is found that strong edges do not always profit kernel estimation, but instead under certain circumstance degrade it, which leads to a new metric to measure the usefulness of image edges in motion deblurring and a gradient selection process to mitigate their possible adverse effect.
- LISA: Reasoning Segmentation via Large Language Model1,078
This work proposes a new segmentation task - reasoning segmentation and presents LISA: large Language Instructed Segmentation Assistant, which inherits the language generation capabilities of multimodal Large Language Models (LLMs) while also possessing the ability to produce segmentation masks.
- Deep Convolutional Neural Network for Image Deconvolution999
This work develops a deep convolutional neural network to capture the characteristics of degradation, establishing the connection between traditional optimization-based schemes and a neural network architecture where a novel, separable structure is introduced as a reliable support for robust deconvolution against artifacts.
- 979
- Structure extraction from texture via relative total variation948
This work proposes new inherent variation and relative total variation measures, which capture the essential difference of these two types of visual forms, and develops an efficient optimization system to extract main structures.
- Exploring Self-Attention for Image Recognition946
This work considers two forms of self-attention, pairwise and patchwise, which generalizes standard dot-product attention and is fundamentally a set operator and strictly more powerful than convolution.
- STD: Sparse-to-Dense 3D Object Detector for Point Cloud866
This work proposes a two-stage 3D object detection framework, named sparse-to-dense 3D Object Detector (STD), and implements a parallel intersection-over-union (IoU) branch to increase awareness of localization accuracy, resulting in further improved performance.
- 840
- Prior Guided Feature Enrichment Network for Few-Shot Segmentation764
The PFENet consists of novel designs of a training-free prior mask generation method that not only retains generalization power but also improves model performance and Feature Enrichment Module (FEM) that overcomes spatial inconsistency by adaptively enriching query features with support features and prior masks.
- 749
- Distilling Knowledge via Knowledge Review717
For the first time in knowledge distillation, cross-stage connection paths are proposed and a new review mechanism is proposed that requires negligible computation overhead, and outperforms other methods on a variety of tasks.
- Image completion with structure propagation671
This paper introduces a novel approach to image completion in which the user manually specifies important missing structure information by extending a few curves or line segments from the known to the unknown regions by adopting the Belief Propagation algorithm to find the optimal patches.
- 617
- PointGroup: Dual-Set Point Grouping for 3D Instance Segmentation610
This paper designs a two-branch network to extract point features and predict semantic labels and offsets, for shifting each point towards its respective instance centroid, and presents PointGroup, a new end-to-end bottom-up architecture specifically focused on better grouping the points by exploring the void space between objects.
- Rolling Guidance Filter590
A new framework to filter images with the complete control of detail smoothing under a scale measure is proposed, based on a rolling guidance implemented in an iterative manner that converges quickly and achieves realtime performance and produces artifact-free results.
- 583
- Detail-Revealing Deep Video Super-Resolution574
This paper shows that proper frame alignment and motion compensation is crucial for achieving high quality results, and proposes a “sub-pixel motion compensation” (SPMC) layer in a CNN framework that can generate visually and quantitatively high-quality results, superior to current state-of-the-arts, without the need of parameter tuning.
- Stratified Transformer for 3D Point Cloud Segmentation564
Stratified Transformer is proposed that is able to capture long-range contexts and demonstrates strong generalization ability and high performance and first-layer point embedding to aggregate local information, which facilitates convergence and boosts performance.
- MAT: Mask-Aware Transformer for Large Hole Image Inpainting539
A novel transformer-based model for large hole inpainting is presented, which unifies the merits of transformers and convolutions to efficiently process high-resolution images.
- VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking524
This paper shows that a fully sparse voxel-based representation works decently for LIDAR 3D object detection and tracking, and proposes VoxelNext, an elegant and efficient framework with no need for sparse-to-dense conversion or NMS post-processing.
- 493
- 481
- Fast Point R-CNN450
This work presents a unified, efficient and effective framework for point-cloud based 3D object detection that achieves state-of-the-arts with a 15FPS detection rate.
- 424
- Mini-Gemini: Mining the Potential of Multi-Modality Vision Language Models415
This work introduces Mini-Gemini, a simple and effective framework enhancing multi-modality Vision Language Models (VLMs), and proposes to utilize an additional visual encoder for high-resolution refinement without increasing the visual token count to enhance visual tokens.
- Parametric Contrastive Learning414
This paper introduces a set of parametric class-wise learnable centers to rebalance from an optimization perspective and demonstrates that PaCo can adaptively enhance the intensity of pushing samples of the same class close as more samples are pulled together with their corresponding centers and benefit hard example learning.
- Improving Calibration for Long-Tailed Recognition414
Motivated by the fact that predicted probability distributions of classes are highly related to the numbers of class instances, this work proposes label-aware smoothing to deal with different degrees of over-confidence for classes and improve classifier learning.
- 414
- 410
- Unifying Voxel-based Representation with Transformer for 3D Object Detection401
The proposed method achieves leading performance in the nuScenes test set for both object detection and the following object tracking task, and surpasses previous work in single- or multi-modality entries.
- Video-P2P: Video Editing with Cross-Attention Control391
This work proposes to first tune a Text-to-Set (T2S) model to complete an approximate inversion and then optimize a shared unconditional embedding to achieve accurate video inversion with a small memory cost, and proves that it is crucial for consistent video editing.
- 385
- SegStereo: Exploiting Semantic Information for Disparity Estimation383
This paper suggests that appropriate incorporation of semantic cues can greatly rectify prediction in commonly-used disparity estimation frameworks and proposes a unified model SegStereo, which employs semantic features from segmentation and introduces semantic softmax loss, which helps improve the prediction accuracy of disparity maps.
- Understanding and Diagnosing Visual Tracking Systems365
A framework by breaking a tracker down into five constituent parts, namely, motion model, feature extractor, observation model, model updater, and ensemble post-processor is proposed, which can provide a solid baseline when conducting controlled experiments for visual tracking research.
- 363
- 363
- VisionZip: Longer is Better but Not Necessary in Vision Language Models362
The proposed VisionZip is a simple yet effective method that selects a set of informative tokens for input to the language model, reducing visual token redundancy and improving efficiency while maintaining model performance.
- Focal Sparse Convolutional Networks for 3D Object Detection326
This paper introduces two new modules to enhance the capability of Sparse CNNs, both are based on making feature sparsity learnable with position-wise importance prediction and shows that spatially learnable sparsity in sparse convolution is essential for sophisticated 3D object detection.
- 310
- LongLoRA: Efficient Fine-tuning of Long-Context Large Language Models297
LongLoRA is presented, an efficient fine-tuning approach that extends the context sizes of pre-trained large language models (LLMs), with limited computation cost, and is compatible with most existing techniques, like Flash-Attention2.
- 295
- Associatively Segmenting Instances and Semantics in Point Clouds292
This paper first introduces a simple and flexible framework to segment instances and semantics in point clouds simultaneously, and proposes two approaches which make the two tasks take advantage of each other, leading to a win-win situation.
- 291
- 278
- 272
- 271
- Spherical Transformer for LiDAR-Based 3D Recognition270
This work study the varying-sparsity distribution of LiDAR points and present SphereFormer to directly aggregate information from dense close points to the sparse distant ones, and design radial window self-attention that partitions the space into multiple non-overlapping narrow and long windows.
- 267
- 263
- Semi-supervised Semantic Segmentation with Directional Context-aware Consistency261
This paper proposes to maintain the context-aware consistency between features of the same identity but with different contexts, making the representations robust to the varying environments, and presents the Directional Contrastive Loss (DC Loss) to accomplish the consistency in a pixel-to-pixel manner.
- 248
- 248
- Deep Automatic Portrait Matting248
An automatic image matting method for portrait images that does not need user interaction is proposed and achieves comparable results with state-of-the-art methods that require specified foreground and background regions or pixels.
- 240
- DSGN: Deep Stereo Geometry Network for 3D Object Detection232
The Deep Stereo Geometry Network (DSGN), for the first time, provides a simple and effective one-stage stereo-based 3D detection pipeline that jointly estimates the depth and detects 3D objects in an end-to-end learning manner.
- 232
- 228
- ResLT: Residual Learning for Long-tailed Recognition219
This work designs the effective residual fusion mechanism – with one main branch optimized to recognize images from all classes, another two residual branches are gradually fused and optimized to enhance images from medium+tail classes and tail classes respectively.
- Hierarchical Point-Edge Interaction Network for Point Cloud Semantic Segmentation219
This work establishes a hierarchical graph framework, where the graph is initialized from a coarse layer and gradually enriched along the point decoding process, and predicts a label to indicate the semantic consistency of the two connected points to enhance point prediction.
- Multi-View Transformer for 3D Visual Grounding210
This paper proposes a Multi-View Transformer (MVT) for 3D visual grounding, in which the position information of the 3D scene under different views are modeled simultaneously and aggregated together in a multi-view space.
- Deep Edge-Aware Filters205
This work makes the attempt to learn a big and important family of edge-aware operators from data based on a deep convolutional neural network with a gradient domain training procedure, which gives rise to a powerful tool to approximate various filters without knowing the original models and implementation details.
- Fully Convolutional Networks for Panoptic Segmentation202
This approach aims to represent and predict foreground things and background stuff in a unified fully convolutional pipeline and outperforms previous box-based and -free models with high efficiency on COCO, Cityscapes, and Mapillary Vistas datasets with single scale input.
- Automatic Portrait Segmentation for Image Stylization202
This paper introduces a new automatic segmentation algorithm dedicated to portraits and describes several portrait filters that exploit the results of this algorithm to generate high‐quality portraits.
- MASA-SR: Matching Acceleration and Spatial Adaptation for Reference-Based Image Super-Resolution199
The MASA network for RefSR is proposed, where two novel modules are designed to address the problems of high-quality correspondence matching and the computational cost by a coarse-to-fine correspondence matching scheme.
- 198
- Hierarchical Dense Correlation Distillation for Few-Shot Segmentation186
This work designs Hierarchically Decoupled Matching Network (HDMNet) mining pixel-level support correlation based on the transformer architecture and proposes a matching module to reduce train-set overfitting and introduce correlation distillation leveraging semantic correspondence from coarse resolution to boost fine-grained segmentation.
- 185
- MuCAN: Multi-correspondence Aggregation Network for Video Super-Resolution185
This work proposes a temporal multi-correspondence aggregation strategy to leverage similar patches across frames, and a cross-scale nonlocal-cor correspondence aggregation scheme to explore self-similarity of images across scales.
- Semi-Parametric Image Synthesis177
A semi-parametric approach to photographic image synthesis from semantic layouts that combines the complementary strengths of parametric and nonparametric techniques is presented.
- LargeKernel3D: Scaling up Kernels in 3D Sparse CNNs174
This work proposes the spatial-wise partition convolution and its large-kernel module, which avoids the optimization and efficiency issues of naive 3D large kernels and shows that large kernels are feasible and essential for 3D visual tasks.
- 164
- Bidirectional Projection Network for Cross Dimension Scene Understanding159
This paper presents a bidirectional projection network (BPNet) for joint 2D and 3D reasoning in an end-to-end manner that achieves top performance on the ScanNetV2 benchmark for both 1D and 2D semantic segmentation.
- Learnable Boundary Guided Adversarial Training158
This paper constrain logits from the robust model that takes adversarial examples as input and makes it similar to those from the clean model fed with corresponding natural data so as to preserve high natural accuracy and benefit model robustness.
- 158
- 154
- 150
- 150
- 149
- 147
- Guided Point Contrastive Learning for Semi-supervised Point Cloud Semantic Segmentation146
This work proposes the guided point contrastive loss to enhance the feature representation and model generalization ability in semi-supervised setting and designs the confidence guidance to ensure high-quality feature learning.
- Video Frame Interpolation with Transformer143
A novel framework is introduced, which takes advantage of Transformer to model long-range pixel correlation among video frames and is equipped with a novel cross-scale window-based attention mechanism, where cross- scale windows interact with each other.
- Situation Recognition with Graph Neural Networks143
A model based on Graph Neural Networks is proposed that allows us to efficiently capture joint dependencies between roles using neural networks defined on a graph and significantly outperforms existing work, as well as multiple baselines.
- 142
- EfficientNeRF - Efficient Neural Radiance Fields141
This paper presents EfficientNeRF as an efficient NeRF-based method to represent 3D scene and synthesize novel-view images and proposes valid and pivotal sampling at the coarse and fine stage, respectively, to significantly improve sampling efficiency.
- 140
- Augmented Feedback in Semantic Segmentation Under Image Level Supervision139
This paper proposes a framework with only image-level supervision that unifies semantic segmentation and object localization with important proposal aggregation and selection modules that greatly reduce the notorious error accumulation problem that commonly arises in weakly supervised learning.
- GAL: Geometric Adversarial Loss for Single-View 3D-Object Reconstruction136
GAL is composed of two terms where the geometric loss ensures consistent shape of reconstructed 3D models close to ground-truth from different viewpoints, and the conditional adversarial loss generates a semantically-meaningful point cloud.
- FocalFormer3D : Focusing on Hard Instance for 3D Object Detection133
This work proposes Hard Instance Probing (HIP), a general pipeline that identifies FN in a multi-stage manner and guides the models to focus on excavating difficult instances, and instantiates this method as FocalFormer3D, a simple yet effective detector that excels at excavating difficult objects and improving prediction recall.
- 131
- 129
- 125
- Generalized Few-shot Semantic Segmentation123
This paper introduces a new benchmark, called Generalized Few-Shot Semantic Segmentation (GFS- Seg), and proposes the Context-Aware Prototype Learning (CAPL) that significantly improves performance by leveraging the co-occurrence prior knowledge from support samples and dynamically enriching contextual information to the classifier, conditioned on the content of each query image.
- Voxel Field Fusion for 3D Object Detection121
This work presents a conceptually simple yet effective framework for cross-modality 3D object detection, named voxel field fusion, and develops mixed augmentor to align feature-variant transformations, which bridges the modality gap in data augmentation.
- 119
- 118
- 117
- 117
- High Quality Entity Segmentation115
CropFormer is the first query-based Transformer architecture that can effectively fuse mask predictions from multiple image views, by learning queries that effectively associate the same entities across the full image and its crop.
- 114
- Video Instance Segmentation with a Propose-Reduce Paradigm113
This work proposes a new paradigm – Propose-Reduce, to generate complete sequences for input videos by a single step, and builds a sequence propagation head on the existing image-level instance segmentation network for long-term propagation.
- 107
- 105
- 104
- 104
- OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation103
The creation of Omni-Adaptive 3D CNNs (OA-CNNs), a family of networks that integrates a lightweight module to greatly enhance the adaptivity of sparse CNNs at minimal computational cost, is proposed.
- Open World Entity Segmentation103
This work proposes a CondInst-like fully-convolutional architecture with two novel modules specifically designed to exploit the class-agnostic and non-overlapping requirements of ES, and investigates the feasibility of convolutional center-based representation to segment things and stuffs in a unified manner.
- Multi-modal and Multi-spectral Registration for Natural Images100
This work designs the descriptor named robust selective normalized normalized cross correlation (RSNCC) to establish dense pixel correspondence in input images and proposes its mathematical param- eterization to make optimization tractable.
- 98
- Combining sketch and tone for pencil drawing production96
A new system to produce pencil drawing from natural images is proposed, which establishes a style that artists use to understand visual data and draw them, and lets the results contain rich and well-ordered lines to vividly express the original scene.
- 96
- 95
- Aggregation via Separation: Boosting Facial Landmark Detector With Semi-Supervised Style Translation95
A new perspective of facial landmark detection is investigated and it is demonstrated it leads to further notable improvement and surprisingly outperforms the fully-supervised model by a large margin.
- Contrast Preserving Decolorization with Perception-Based Quality Metrics94
This paper proposes an optimization framework aiming at maximally preserving color contrast, and employs a bimodal objective function to alleviate the restrictive order constraint for color mapping, and develops an efficient solver that allows for automatic selection of suitable grayscales based on global contrast constraints.
- PFENet++: Boosting Few-Shot Semantic Segmentation With the Noise-Filtered Context-Aware Prior Mask93
This work proposes the Context-aware Prior Mask (CAPM) that leverages additional nearby semantic cues for better locating the objects in query images and takes one step further by incorporating a lightweight Noise Suppression Module (NSM) to screen out the unnecessary responses, yielding high-quality masks for providing the prior knowledge.
- GeoNet++: Iterative Geometric Neural Network with Edge-Aware Refinement for Joint Depth and Surface Normal Estimation93
A new 3D geometric metric (3DGM) for evaluating depth prediction in 3D, which measures whether the predicted depth can reconstruct high quality 3D surface normals and is a more natural metric for many 3D application domains.
- Generalized Parametric Contrastive Learning91
This analysis demonstrates that GPaCo/PaCo can adaptively enhance the intensity of pushing samples of the same class close as more samples are pulled together with their corresponding centers and benefit hard example learning.
- Visual Question Answering with Question Representation Update (QRU)88
This model contains several reasoning layers, exploiting complex visual relations in the visual question answering (VQA) task, end-to-end trainable through back-propagation, where its weights are initialized using pre-trained convolutional neural network (CNN) and gated recurrent unit (GRU).
- 88
- 87
- VCNet: A Robust Approach to Blind Image Inpainting86
This paper proposes a two-stage visual consistency network (VCN), meant to estimate where to fill (via masks) and generate what to fill and repairs these estimated missing regions using a new spatial normalization, enabling VCN to be robust to the mask prediction errors.
- InternVLA-M1: A Spatially Guided Vision-Language-Action Framework for Generalist Robot Policy85
InternVLA-M1 outperforms its variant without spatial guidance in long-horizon reasoning-intensive scenarios, and surpassed existing works by over 10%.
- 83
- A Segmentation Based Variational Model for Accurate Optical Flow Estimation83
A new optical flow scheme containing three phases that not only produces quantitatively accurate optical flow estimates but also preserves sharp motion boundaries, which makes the optical flow result usable in a number of computer vision applications, such as image/video segmentation and editing.
- 82
- Rethinking Out-of-distribution (OOD) Detection: Masked Image Modeling is All You Need81
This work takes Masked Image Modeling as a pretext task for OOD detection framework (MOOD) and finds that reconstruction-based pretext tasks have the potential to provide a generally applicable and efficacious prior, which benefits the model in learning intrinsic data distributions of the ID dataset.
- A sparse control model for image and video editing80
This work discovers inherent limitation of existing methods for their implicit requirement on where and how the strokes are drawn, and presents a new system that is principled on minimizing the amount of work put in user interaction.
- Stitcher: Feedback-driven Data Provider for Object Detection76
Stitcher is presented, a feedback-driven data provider, which aims to train object detectors in a balanced way, and steadily improves performance by a large margin in all settings, especially for small objects, with nearly no additional computation in both training and testing stages.
- Understanding Imbalanced Semantic Segmentation Through Neural Collapse75
It is pointed out that semantic segmentation naturally brings contextual correlation and imbalanced distribution among classes, which breaks the equiangular and maximally separated structure of neural collapse for both feature centers and classifiers, and a regularizer on feature centers is in-troduce to preserve these advantages.
- Image Synthesis via Semantic Composition75
A novel approach to synthesize realistic images based on their semantic layouts hypothesizes that for objects with similar appearance, they share similar representation, and proposes a dynamic weighted network constructed by spatially conditional computation.
- Bayesian Correction of Image Intensity with Spatial Consideration75
A novel approach to recover a high-quality image by exploiting the tradeoff between exposure time and motion blur, which considers color statistics and spatial constraints simultaneously, by using only two defective input images.
- Multi-Scale Aligned Distillation for Low-Resolution Detection74
The idea of spatially aligning feature maps between models of varying input resolutions by shifting feature pyramid position is explored and crossed feature-level fusion is proposed to dynamically fuse teacher’s multi-resolution features to guide the student better.
- Efficient Non-Consecutive Feature Tracking for Robust Structure-From-Motion74
This paper proposes an efficient non-consecutive feature tracking framework to match interrupted tracks distributed in different subsequences or even in different videos, and contributes an effective segment-based coarse-to-fine SfM algorithm for robustly handling large data sets.
- 72
- Jigsaw Clustering for Unsupervised Visual Representation Learning71
A new jigsaw clustering pretext task is proposed, which only needs to forward each training batch itself, and reduces the training cost, and outperforms previous single-batch based ones by a large margin.
- 69
- 69
- Real-time contrast preserving decolorization69
A very fast and yet effective decolorization approach is proposed that is borne out by a new quantitative metric as well as qualitative comparisons with state-of-the-art methods.
- 68
- Mask-Attention-Free Transformer for 3D Instance Segmentation67
This work abandons the mask attention design and resorts to an auxiliary center regression task instead, which effectively overcome the low-recall issue and perform cross-attention by imposing positional prior.
- 67
- 67
- 66
- Fully Convolutional Networks for Panoptic Segmentation with Point-based Supervision65
A conceptually simple, strong, and efficient framework for fully- and weakly-supervised panoptic segmentation, called Panoptic FCN, which encodes each object instance or stuff category with the proposed kernel generator and produces the prediction by convolving the high-resolution feature directly.
- 64
- High Quality Segmentation for Ultra High-resolution Images63
The Continuous Refinement Model (CRM) continuously aligns the feature map with the refinement target and aggregates features to reconstruct these image details and shows its significant generalization ability to fill the resolution gap between low-resolution training images and ultra high-resolution testing ones.
- 57
- 55
- 55
- Stereo Matching: An Outlier Confidence Approach55
The Outlier Confidence (OC) which dynamically measures how likely one pixel is occluded is introduced which locally infers the possible disparity values for the outlier pixels using a reliable color sample refinement scheme.
- 53
- A Unified Query-based Paradigm for Point Cloud Understanding53
Extensive experimental results on various 3D tasks show that EQ-Paradigm in tandem with Q-Net is a general and effective pipeline, which enables flexible collaboration of backbones and heads and boosts performance of state-of-the-art methods.
- Inverse Kernels for Fast Spatial Deconvolution50
This work proposes an efficient spatial deconvolution method that can incorporate sparse priors to suppress noise and visual artifacts and is based on estimating inverse kernels that are decomposed into a series of 1D kernels.
- 50
- 49
- Dynamic Divide-and-Conquer Adversarial Training for Robust Semantic Segmentation49
A dynamic divide-and-conquer adversarial training (DDC-AT) strategy to enhance the defense effect, by set-ting additional branches in the target model during training, and dealing with pixels with diverse properties to-wards adversarial perturbation.
- 49
- Ref-NPR: Reference-Based Non-Photorealistic Radiance Fields for Controllable Scene Stylization48
The experimental results demonstrate that Ref-NPR out-performs existing scene and video stylization methods regarding visual quality and semantic correspondence.
- Break Ames room illusion48
A solution based on small-scale defocus blur inherent in optical lens is proposed and it incorporates a matching prior with the newly constructed edgelet dataset using a non-local scheme, and includes semantic depth order cues for physically based inference.
- DSGN++: Exploiting Visual-Spatial Relation for Stereo-Based 3D Detectors46
This work revisits the prior stereo detector DSGN for its stereo volume construction ways for representing both 3D geometry and semantics and proposes the advanced version, DSGN++, aiming to enhance effective information flow throughout the 2D-to-3D pipeline.
- Efficient Non-consecutive Feature Tracking for Structure-from-Motion45
This paper addresses the non-consecutive feature point tracking problem and proposes an effective method to match interrupted tracks and results show that the method notably improves SfM.
- 44
- 44
- Facelet-Bank for Fast Portrait Manipulation43
A model based on an end-to-end convolutional neural network that supports fast inference, edit-effect control, and quick partial-model update is proposed that can handle a wide range of expressions, accessories, and makeup effects.
- 42
- 40
- Attentive Normalization for Conditional Image Generation39
This paper characterize long-range dependence with attentive normalization (AN), which is an extension to traditional instance normalization, which enhances consistency between distant regions with semantic correspondence.
- 39
- Consistent Binocular Depth and Scene Flow with Chained Temporal Profiles38
The main contributions include development of motion trajectories, which robustly link frame correspondences in a voting manner, rejection of depth/motion outliers through temporal robust regression, novel edge occurrence map estimation, and introduction of anisotropic smoothing priors for proper regularization.
- 38
- 35
- PointINS: Point-Based Instance Segmentation34
This paper proposes an instance-aware convolution that decomposes this mask representation learning task into two tractable modules as instance- aware weights and instance-agnostic features, improving mask learning efficiency by avoiding employing several independent convolutions.
- 34
- 34
- 32
- 30
- 30
- 29
- 28
- 27
- 27
- 27
- 26
- 25
- 25
- 25
- 25
- 24
- 23
- 22
- 21
- 21
- Dense scattering layer removal21
This work proposes a new model, together with advanced optimization, to separate a thick scattering media layer from a single natural image, and introduces non-local structure-aware regularization to properly constrain transmission estimation without introducing the halo artifacts.
- 20
- 20
- 20
- 19
- 19
- 19
- 18
- 18
- 18
- 17
- 17
- 16
- 16
- 16
- 16
- 16
- 16
- 15
- 15
- 14
- 14
- 14
- 13
- 13
- 13
- 13
- 12
- 12
- 12
- 11
- 10
- 10
- 10
- 9
- 9
- 9
- 9
- 8
- 8
- 8
- 7
- Empowering Visual Creativity: A Vision-Language Assistant to Image Editing Recommendations6
This paper introduces Creativity-Vision Language Assistant(Creativity-VLA), a multimodal framework designed specifically for edit-instruction generation that aims to automatically generate diverse creative editing instructions from an input image and a simple prompt representing the users' under-specified editing purpose.
- 6
- General Adversarial Defense Against Black-box Attacks via Pixel Level and Feature Level Distribution Alignments6
This paper uses Deep Generative Networks with a novel training mechanism to eliminate the distribution gap between adversarial and clean samples in feature space of the target DNNs, and proposes a more effective pixel level training constraint to make this achievable, thus enhancing robustness on adversarial samples.
- 6
- 5
- 5
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 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.