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
- Yiwei JiangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works174 from public data
- Depth Matters: Influence of Depth Cues on Visual Saliency297
This work collects a large human eye fixation database compiled from a pool of 600 2D-vs-3D image pairs viewed by 80 subjects, where the depth information is directly provided by the Kinect camera and the eye tracking data are captured in both 2D and 3D free-viewing experiments.
- Collaborative Perception in Autonomous Driving: Methods, Datasets, and Challenges281
This article systematically summarize the collaborative perception methods for ideal scenarios and real-world issues, and highlights gaps and overlooked challenges between current academic research andreal-world applications.
- 199
- Partial Multi-Label Learning by Low-Rank and Sparse Decomposition145
This paper utilizes the low-rank and sparse decomposition scheme and proposes a novel Partial Multi-label Learning by Low-Rank and Sparse decomposition (PML-LRS) approach, which reformulates the observed label set into a label matrix, and decomposes it into a groundtruth label matrix and an irrelevant label matrix.
- GM-PLL: Graph Matching Based Partial Label Learning92
This paper proposes a novel Graph Matching based Partial Label Learning (GM-PLL) framework, where graph Matching scheme is incorporated owing to its excellent capability of exploiting the instance and label relationship, and proposes a relaxed matching prediction model, which can improve the prediction accuracy via GM strategy.
- MSO: Multi-Feature Space Joint Optimization Network for RGB-Infrared Person Re-Identification82
This work proposes an edge features enhancement module to enhance the modality-sharable features in each single-modality space and designs a perceptual edge features (PEF) loss after the edge fusion strategy analysis, which markedly improves the network's performance.
- 78
- 77
- Recurrent convolutional network for video-based smoke detection65
This work proposes a novel video-based smoke detection method via Recurrent Neural Networks (RNNs), which first captures the space and motion context information by using deep convolutional motion-space networks and effectively train the smoke model.
- 64
- Salient Object Detection via Low-Rank and Structured Sparse Matrix Decomposition62
In this model, a tree-structured sparsity-inducing norm regularization is firstly introduced to provide a hierarchical description of the image structure to ensure the completeness of the extracted salient object, and high-level priors are integrated to guide the matrix decomposition and enhance the saliency detection.
- 52
- 51
- 47
- Age Estimation Based on Convolutional Neural Network42
The proposed method uses Convolutional Neural Network CNN to extract facial features and divides different ages into 13 groups and uses the Support Vector Machine SVM classifier to perform the classification.
- 40
- 31
- 30
- 27
- Multi-human Parsing with a Graph-based Generative Adversarial Model27
A novel multi- human parsing model named MH-Parser is introduced, which uses a graph-based generative adversarial model to address the challenges of close-person interaction and occlusion in multi-human parsing.
- 26
- Joint Graph Learning and Matching for Semantic Feature Correspondence25
The graph patterns learnt by the proposed GLAM model are validated to be able to remarkably enhance previous deep graph matching methods by replacing their handcrafted graph structures with the learnt ones.
- Vehicle detection and classification based on convolutional neural network24
This paper extracts visual features from the activation of a deep convolutional network, large-scale sparse learning and other distinguishing features in order to compare their accuracy.
- Multi-task clustering ELM for VIS-NIR cross-modal feature learning21
A multi-task clustering based ELM is proposed to improve the recognition accuracy by learning the shared information between tasks to reduce the obtained cross-modality features.
- 21
- Beyond tag relevance21
This paper presents an integrated framework for tag saliency ranking which combines both visual attention model and multi-instance learning algorithm to investigate the saliencyranking order information of tags with respect to the given image.
- HERA20
A novel PLL approach named HERA, which simultaneously incorporates the HeterogEneous Loss and the SpaRse and Low-rAnk procedure to estimate the labeling confidence for each instance while training the desired model, is proposed.
- Towards Real World Human Parsing: Multiple-Human Parsing in the Wild.20
A novel Multiple-Human Parser (MH-Parser) is proposed, which considers both the global context and local cues for each person in the parsing process and is demonstrated to outperform the naive "detect-and-parse" approach by a large margin.
- RL-SeqISP: Reinforcement Learning-Based Sequential Optimization for Image Signal Processing18
A novel Sequential ISP parameter optimization model, called the RL-SeqISP model, which utilizes deep reinforcement learning to jointly optimize all ISP parameters for a variety of imaging applications and can progressively enhance image quality by seamlessly integrating information from both the image feature space and the parameter space.
- Dual Low-Rank Pursuit: Learning Salient Features for Saliency Detection18
This paper considers a specific saliency detection problem of predicting human eye fixations when they freely view natural images, and proposes a novel dual low-rank pursuit (DLRP) method that learns saliency-aware feature transformations by utilizing available supervision information and constructs discriminative bases for effectively detecting human fixation points under the popular low- rank and sparsity-pursuit framework.
- Dynamic Dictionary Learning for Remote Sensing Image Segmentation17
A dynamic dictionary learning framework that explicitly models class ID embeddings through iterative refinement is introduced, ensuring compact intra-class distributions while maximizing inter-class separability and a contrastive constraint is applied to the dictionary space.
- 17
- End-to-End Text-to-Image Synthesis with Spatial Constrains17
This article proposes a novel end-to-end approach for text- to-image synthesis with spatial constraints by mining object spatial location and shape information by fusing text semantic and spatial information into a synthesis module and jointly fine-tuning them with multi-scale semantic layouts generated.
- Hierarchical sparse representation based Multi-Instance Semi-Supervised Learning with application to image categorization17
A novel instance disambiguation strategy based on sparse representation that can identify the instance confidence value in both positive and unlabeled bags more effectively is proposed.
- Exposure Fusion Based on Shift-Invariant Discrete Wavelet Transform *16
A novel way to get the chrominance information of the scene, and the saturation of the fused image can be adjusted using one user-controlled parameter is presented, and SIDWTBEF can give comparative results compared to other shift dependent exposure fusion methods.
- Information Theoretic Metrics in Shot Boundary Detection16
A new set of metrics, information theoretic metrics, to quantitatively measure the changes between frames, includes image entropy difference, joint entropy, conditional entropy, mutual information and divergence, which can be used to cut detection.
- HSS-GCN: A Hierarchical Spatial Structural Graph Convolutional Network for Vehicle Re-identification15
A hierarchical spatial structural graph convolutional network (HSS-GCN) for vehicle Re-ID is proposed, in which a hierarchical spatial structure graph is constructed with the global region and local regions as nodes and a two-hierarchy relationship as edges, and later learning discriminative structure features with a GCN module under the constraints of metric learning.
- Adapting Vision Foundation Models for Robust Cloud Segmentation in Remote Sensing Images14
This article presents a parameter-efficient adaptive approach, termed Cloud-Adapter, designed to enhance the accuracy and robustness of cloud segmentation, which leverages a VFM pretrained on general domain data, which remains frozen, eliminating the need for additional training.
- Semi-supervised dual low-rank feature mapping for multi-label image annotation14
A dual low-rank regularized multi-label learning model under a graph regularized semi-supervised learning framework, which can effectively capture the label correlations in the learned feature space, and enforce the label matrix be self-recovered in label space as well.
- Video-Based Illumination Estimation14
A video-based illumination-estimation algorithm that takes advantage of the related information between adjacent frames and outperforms the original single-frame methods on which it is based.
- Graph regularized low-rank feature mapping for multi-label learning with application to image annotation13
A novel graph regularized low-rank feature mapping for image annotation under semi-supervised multi-label learning framework that can explicitly take into account the local geometric structure on both labeled and unlabeled images.
- Generated and Pseudo Content guided Prototype Refinement for Few-shot Point Cloud Segmentation12
A novel framework called G enerated and P seudo C ontent guided P rototype R efinement (GPCPR), which explicitly leverages LLM-generated content and reliable query context to enhance prototype quality and introduces a dual-distillation regularization term, enabling knowledge transfer between early-stage entities (prototypes or pseudo predictions) and their deeper counterparts to enhance refinement.
- 12
- Object detection by crossing relational reasoning based on graph neural network12
This work proposes a novel object detection framework that fully explores the relationships across visual feature space and label embedding space to facilitate the proposal classification in object detection and develops a graph convolutional module to perform relational reasoning on the graph.
- 12
- Co-saliency Detection with Graph Matching12
A novel graph-matching based model is proposed to integrate the visual appearance, saliency coherence, and spatial structural continuity for detecting co-saliency collaboratively and outperforms state-of-the-art baselines significantly.
- 11
- 11
- Saliency ranker: A new salient object detection method11
A novel learning-to-rank model with a limited number of training data is developed, which combines the strength of cost-sensitive label ranking methods with the power of low-rank matrix recovery theories.
- 11
- Colour constancy based on texture similarity for natural images11
Weibull parameterisation is used to identify the texture characteristics of colour images and the results show that this new approach outperforms current state-of-the-art single algorithms, as well as some combined algorithms.
- 11
- FND-SLAM: A SLAM System Using Feature Points and NeRF in Dynamic Environments Based on RGB-D Sensors10
- GLAN: A graph-based linear assignment network10
A learnable linear assignment solver based on deep graph networks that outperforms state-of-the-art baselines and achieves consistently high accuracy with the increment of the problem size is proposed.
- SSR-Net: A Spatial Structural Relation Network for Vehicle Re-identification10
A Spatial Structural Relation Network (SSR-Net) is proposed that explores the above-mentioned two kinds of relations simultaneously to learn more discriminative features by modeling the spatial structure information and global context information.
- Distance-Preserving Embedding Adaptive Bipartite Graph Multi-View Learning with Application to Multi-Label Classification10
This article proposes an anchor-based bipartite graph embedding approach to accelerate the learning process and designs a concise yet effective feature collinearity guided feature selection scheme to learn tight multi-label classifiers.
- 10
- Branching Path Following for Graph Matching10
A novel path following strategy, named branching path following (BPF), is proposed, which consequently improves graph matching performance and achieves remarkable improvement in matching accuracy and outperforms other algorithms.
- Practical anonymity models on protecting private weighted graphs10
A general anonymity model against weight-related attacks, and a new utility metric based on spectral graph theory is introduced, which is effective in terms of efficiency and data utility preservation.
- Towards relevance and saliency ranking of image tags10
An adaptive all-season tag ranking algorithm which can handle the images with and without distinct object(s) using different tag ranking strategies is presented.
- Clicking Matters: Towards Interactive Human Parsing9
This work focuses on Interactive Human Parsing, which aims to segment a human image into multiple human body parts with guidance from users’ interactions and proposes a semantic-perceiving loss to augment the training, which can effectively exploit the semantic relationships of clicks for better optimization.
- 9
- 9
- Seeing Crucial Parts: Vehicle Model Verification via a Discriminative Representation Model8
This article introduces a simple yet powerful deep model—the enforced intra-class alignment network (EIA-Net)—which can learn a more discriminative image representation by localizing key vehicle parts and jointly incorporating two distance metrics: vehicle-level embedding and vehicle-part-sensitive embedding.
- 8
- Pedestrian attribute recognition based on attribute correlation7
A Cross Attribute and Feature Network (CAFN) that fully exploits the correlations between any pair of attributes for the pedestrian attribute recognition to tackle challenges of class imbalance and partial occlusion is proposed.
- 7
- 7
- 6
- 6
- Deep spatio-frequency saliency detection6
A Spatio-Frequency Network (SFNet) is proposed that exploits spatio-frequency clues to effectively enlarge the receptive fields of CNN layers and more importantly, strengthen their spatial discrimination for better saliency detection.
- Learning part-alignment feature for person re-identification with spatial-temporal-based re-ranking method6
A network to learn powerful features combining global features and local-alignment features for person re-identification and a novel re-ranking approach is proposed to refine the rough initial rank list by exploiting the spatial-temporal information.
- Robust Semi-supervised Multi-label Learning by Triple Low-Rank Regularization6
A linear self-representative model is firstly introduced to recover the possibly noisy label matrix by exploiting the label correlations, and a low-rank pairwise similarity matrix is developed to capture the global relationships among labeled and unlabeled samples by taking advantage of Low-Rank Representation.
- 6
- General structured sparse learning for human facial age estimation6
The learning problem is formulated as a structured sparse multi-class classification model regularized by the structured sparsity, which encodes the ordinal relationships among different age labels, and ensures that the samples with similar age labels are close to each other in the feature space.
- 6
- From sample selection to model update: A robust online visual tracking algorithm against drifting6
This paper proposes an online tracking algorithm that employs a confidence combinatorial map model that describes the structure of the object and demonstrates that the proposed algorithm outperforms other comparative algorithms and gives a very good performance.
- 6
- Knowledge Transfer and Domain Adaptation for Fine-Grained Remote Sensing Image Segmentation5
This work introduces a novel end-to-end learning paradigm combining knowledge guidance with domain refinement to enhance performance and highlights the potential of combining knowledge transfer and domain adaptation to overcome domain-related challenges and data limitations.
- 5
- 5
- Symmetry-aware graph matching5
A new framework for symmetry-aware graph matching in the context of image matching is proposed, and for the first time symmetry constraints are introduced into graph matching.
- An error-tolerant approximate matching algorithm for labeled combinatorial maps5
This paper addresses the problem of inexact matching between labeled combinatorial maps, and aims to find a rapid algorithm for measuring distance between maps by proposing a distance measure based on joint-trees and an approximate approach.
- Adaptive all-season image tag ranking by saliency-driven image pre-classification5
An adaptive all-season tag ranking algorithm which can handle the images with and without distinct object(s) using different tag ranking strategies is presented.
- Mixture of Global and Local Experts With Diffusion Transformer for Controllable Face Generation4
Face-MoGLE significantly outperforms state-of-the-art controllable face generation methods in both unimodal and multimodal settings, achieving superior image quality, semantic alignment, aesthetic preference, and zero-shot generalization.
- 4
- Deep Probabilistic Graph Matching4
A deep learning-based graph matching framework that works for the original QAP without compromising on the matching constraints is proposed, and the probabilistic solver refines the estimated assignments in an iterative manner to impose both discrete and one-to-one matching constraints.
- UNAGI: Unified neighbor-aware graph neural network for multi-view clustering4
A novel framework that seamlessly merges the optimization of the graph topology and sample representations through a differentiable graph adapter, which enables a unified training paradigm for multi-vIew clustering (UNAGI).
- 4
- SMM: Self-supervised Multi-Illumination Color Constancy Model with Multiple Pretext Tasks4
A self-supervised multi-illumination color constancy model that leverages multiple pretext tasks to fully explore lighting color contextual information and inherent color information without using any manual annotations is proposed.
- 4
- Salient object detection with side information4
This work adopts the side information such as image-level tags labeled on the objects in images to improve the saliency detection results and extends the global average pooling (GAP) to predict the salient object in complex images and use it as a layer in the deep convolutional network.
- Improving person re-identification via attribute-identity representation and visual attention mechanism4
Improve the appearance of a person by improving feature descriptors that shed light on a combination framework by fusing the attribute-identity discrimination network with the person discrimination network based on the visual attention mechanism.
- Graph-based visual odometry for VSLAM4
A novel graph-based method that aims to improve matching accuracy by fully exploiting the structure information and is compared with several state-of-the-art visual simultaneous localization and mapping algorithms on three datasets.
- A framework with a multi-task CNN model joint with a re-ranking method for vehicle re-identification4
This work proposes a new framework with a multi-task CNN model and a ranking optimization method to tackle the re-id task and combines the strengths of the two models, which can get a more discriminative feature of vehicle image.
- A Novel Emotional Saliency Map to Model Emotional Attention Mechanism4
This work presents an intuitive emotional saliency map computation method by calculating Minkowski-norm of pixel’s isolated saliency value and multi-scale local contrast information in color emotion space and shows that the proposed model can outperform some current state-of-the-art visual saliency maps.
- Combining visual attention model with multi-instance learning for tag ranking4
A novel tag saliency ranking scheme, which aims to automatically rank the tags associated with a given image according to their saliency to the image content, and combines both visual attention model and multi-instance learning to investigate the saliencyranking order information of tags with respect to the given image.
- An Adaptive Tone Mapping Method for Displaying High Dynamic Range Images4
A Local Adaptive Bilateral Filter (LABF) method having threefold improvement over the bilateral filter based method to imitate the human response to light, and an improved center/surround method to enhance visibility in the highlight or relatively darker areas further.
- Perception-Oriented Prominent Region Detection in Video Sequences Using Fuzzy Inference Neural Network4
A modified Fuzzy Inference Neural Network is used to detect prominent regions in video sequences, due to its simple structure and superior performance for automatic fuzzy rules extraction.
- Toward Stable Semi-Supervised Remote Sensing Segmentation via Co-Guidance and Co-Fusion3
This work constructs a heterogeneous dual-student architecture comprising two distinct ViT-based vision foundation models initialized with the weights of pretrained CLIP and DINOv3 to mitigate error accumulation and pseudo-label drift, and introduces a semantic co-guidance mechanism that utilizes text embeddings and learnable queries to provide explicit and implicit class-level guidance.
- 3
- Joint Homophily and Heterophily Relational Knowledge Distillation for Efficient and Compact 3D Object Detection3
A novel methodology termed Joint Homophily and Heterophily Relational Knowledge Distillation (H2RKD) is proposed to distill robust relational knowledge in point clouds, thereby enhancing intra-object similarity and refining inter-object distinction.
- Fine-Grained Semantic Image Synthesis with Object-Attention Generative Adversarial Network3
The proposed OA-GAN is a novel object-attention generative adversarial network that allows attention-driven, multi-fusion refinement for fine-grained semantic image synthesis and significantly outperforms the state-of-the-art methods.
- 3
- HERA: Partial Label Learning by Combining Heterogeneous Loss with Sparse and Low-Rank Regularization3
- Stabilizing video facial landmark detection and tracking via global and local filtering3
The main contribution within the solution is a novel global and local filtering strategy, which guarantees the robustness of global whole facial shape tracking and the adaptivity of local facial parts tracking.
- 3
- Realtime Human Segmentation in Video3
A new real-time video-based human segmentation framework which is designed for the single person from videos to produces smoothing, accurate and fast human segmentsation results.
- A novel hypergraph matching algorithm based on tensor refining3
This work proposes a novel hypergraph matching algorithm based on tensor refining, accompanied with an alternative adjustment method to accelerate the convergence, and makes a comparison between the proposed approach and several outstanding matching algorithms on three commonly used benchmarks.
- Depth Information Fused Salient Object Detection3
Several state-of-the-art saliency detection models can be utilized to generate 2D salient maps, which can be fused with the depth map to detect the salient object in a given image.
- Contextualizing Tag Ranking and Saliency Detection for Social Images3
This paper investigates how to iteratively and mutually boost tag ranking and saliency detection by taking the outputs from one task as the context of the other one and proposes an integrated framework for tag saliency ranking which combines both visual attention model and multi-instance learning.
- 3
- A unified supervised codebook learning framework for classification3
A discriminative visual dictionary learning method for boosting the classification performance by relaxing the hard binary constraints to soft nonnegative ones and a multiplicative nonnegative update procedure is proposed to optimize the objective function along with theoretic convergence proof.
- 3
- Shot Type Classification in Sports Video Using Fuzzy Information Granular3
A new method for classifying shot type in sports video using fuzzy information granular, which can adequately capture the uncertainty or ambiguity of scales of a shot.
- 2
- 2
- 2
- Redundant Label Learning via Subspace Representation and Global Disambiguation2
A novel feature SubspacE Representation and label Global DisambiguatIOn (SERGIO) approach, which improves the generalization ability of the learning system from the perspective of both feature space and label space and introduces a label confidence matrix and constrain it with trace-norm regularization simultaneously.
- MFAD: A Multi-modality Face Anti-spoofing Dataset2
The influences of different modalities on the performance of face anti-spoofing are studied by extensively experimenting with different combinations of modalities with an end-to-end deep learning model, and it is found that the HDR information contributes most among differentmodalities.
- 2
- Partial label learning via low rank representation and label propagation2
A novel two-stage approach to learn from the partial label instances that obtains a weighted graph by seeking low rank and sparse matrix in the feature space and the every entry is non-negative and disambiguates the candidate label set by the label propagation.
- Multiple path exploration for graph matching2
A novel algorithm to bypass this pitfall of singular point discovering by checking the smoothness of the path and subsequently explore multiple smooth curves at detected points for better matching results is designed.
- Salient object detection of social images based on semantic tag context2
This paper proposes a new paradigm on salient object detection, which aims at producing more reliable results by mining the context information from user's annotated tags by taking account of both tag contexts and appearance cues.
- Facial Age Estimation Based on Structured Low-rank Representation2
An algorithm based on structured, low- rank representation for facial age estimation based on the constraint of the classwise block-diagonal structure to promote discrimination of representations for robust recognition is presented.
- Personalized image recommendation and retrieval via latent SVM based model2
A predictive framework based on the latent SVM model to retrieve the most relevant images from the dataset at an individual user level is developed, which models the relations between scene-level features and the global- level features that influence it in a globally optimal way.
- 2
- 1
- 1
- 1
- 1
- Enhancing Multimedia Applications by Removing Dynamic Objects in Neural Radiance Fields1
A novel framework to exclude dynamic interference from NeRF scenes is presented, enhancing the practicability in multimedia applications and significantly out-performs existing methods in terms of appearance metrics for the task of removing dynamic objects from scenes.
- 1
- 1
- 1
- Saliency Aggregation: Multifeature and Neighbor Based Salient Region Detection for Social Images1
A deep convolution neural network is built, which considers both appearance features and tag features, and tag neighbor and appearance neighbor based saliency aggregation terms are added to the saliency model to enhance salient regions.
- 1
- 1
- NG-GS: NeRF-Guided 3D Gaussian Splatting Segmentation–
NG-GS is introduced, a novel framework for high-quality object segmentation in 3DGS that explicitly addresses boundary discretization and achieves state-of-the-art performance on NVOS, LERF-OVS, and ScanNet benchmarks.
- –
- –
- –
- –
- –
- –
- –
- –
- Adaptive spatial-temporal graph ODE networks for traffic flow forecasting–
Adaptive Spatial-temporal Graph ODE Networks (ASTGODE) is proposed, which leverage Neural ODEs to reformulate discrete spatial-temporal convolutions into a continuous modeling paradigm, and incorporates multi-modal graph convolutions to capture heterogeneous spatial dependencies.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Unsupervised Person Re-identification Based on Clustering and Domain-Invariant Network–
A novel domain adaptation framework for unsupervised person Re-ID is proposed, where target data are clustered and selected to add relative reliable supervised information for target domain and a novel domain adaptive network is designed to decompose the representations to person-related and domain-related part.
- –
- –
- A general method for sensitive identification detection in the terrorist video–
A general method for sensitive identification detection in the terrorist video under the cluttered scenes using a Haar-like wavelet representation and a novel gradient histogram feature to capture the structural similarities between the target images.
- Joint Tree of Combinatorial Maps–
The Joint Tree of combinatorial maps is defined, and it is proved it can be used to decide of map isomorphism, and the map matching problem is relaxed to the Joint Tree matching problem, which can be solved in polynomial time.
- –
- –
- –
- –
- Visual Morphing Based on the Compressed Domain–
Two image morphing algorithms in the discrete cosine transform (DCT) domain are proposed and a morphing method following the idea of fluid simulation is proposed, which has demonstrated the convenience, efficiency and accuracy of the proposed algorithm.
- Adaptive Non-linear Intensity Mapping Based Salient Region Extraction–
This paper proposes a robust approach to extracting the salient region based on bottom-up visual attention that is effective and can characterize the human perception well, and proposes a constructive approach for rendering an image by a non-linear intensity mapping.
- –
- –
- A Visual Inpainting Method Based on the Compressed Domain–
A visual inpainting model in the discrete cosine transform (DCT) domain is proposed, where DCT coefficients of the non-inpainting blocks are utilized to get block features, and those block features are propagated to the inPainting region iteratively.
- An event detection framework in video sequences based on hierarchic event structure perception–
The main purpose of this paper is not to detect event in special domain, but to construct general event detection framework in perceptual manner and to provide meaningful unit in different semantic granularity.
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 2026-10-15. 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.