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
- Yong Hai YuSuggested from co-authorship
- Shuicheng YanSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works150 from public data
- Robust Recovery of Subspace Structures by Low-Rank Representation3,594
It is shown that the convex program associated with LRR solves the subspace clustering problem in the following sense: When the data is clean, LRR exactly recovers the true subspace structures; when the data are contaminated by outliers, it is proved that under certain conditions LRR can exactly recover the row space of the original data.
- Robust Subspace Segmentation by Low-Rank Representation1,756
Both theoretical and experimental results show that low-rank representation is a promising tool for subspace segmentation from corrupted data.
- 679
- 566
- 220
- 217
- 199
- 176
- Spatio-temporal convolutional features with nested LSTM for facial expression recognition161
A novel end-to-end architecture termed Spatio-Temporal Convolutional features with Nested LSTM (STC-NLSTM), which learns the muti-level appearance features and temporal dynamics of facial expressions in a joint fashion is proposed.
- 161
- 89
- 88
- Active Subspace: Toward Scalable Low-Rank Learning86
This work revisits the classic mechanism of low-rank matrix factorization, based on which an active subspace algorithm is presented for efficiently solving nuclear norm regularized optimization problems by transforming large-scale NNROPs into small-scale problems.
- Joint Label Prediction Based Semi-Supervised Adaptive Concept Factorization for Robust Data Representation82
A joint label prediction based Robust Semi-Supervised Adaptive Concept Factorization (RS2ACF) framework that incorporates the joint neighborhood reconstruction error over the new representations and predicted labels of both labeled and unlabeled data, so the manifold structures can be preserved explicitly and adaptively in the representation space and label space at the same time.
- Deeper cascaded peak-piloted network for weak expression recognition80
Experimental results on two popular facial expression databases, CK+$$+ and Oulu-CASIA, show the superiority of the proposed Deeper Cascaded Peak-piloted Network over state-of-the-art methods.
- Blessing of Dimensionality: Recovering Mixture Data via Dictionary Pursuit74
It is mathematically proved that LRR can weaken the dependence on the row-coherence, provided that the dictionary is well-conditioned and has a rank of not too high, which provides some elementary principles for dictionary learning and naturally lead to a practical algorithm for recovering mixture data.
- 73
- Learning image compressed sensing with sub-pixel convolutional generative adversarial network69
The proposed novel sub-pixel convolutional generative adversarial network (GAN) to learn compressed sensing reconstruction of images is superior to some state-of-the-art deep learning based and iterative optimization based algorithms, in terms of both time complexity and reconstruction quality.
- 68
- 64
- 60
- Scalable Block-Diagonal Locality-Constrained Projective Dictionary Learning59
This work proposes a novel structured discriminative block-diagonal dictionary learning method, referred to as scalable Locality-Constrained Projective Dictionary Learning (LC-PDL), for efficient representation and classification and incorporates a locality constraint of atoms into the DL procedures to keep local information and obtain the codes of samples over each class separately.
- Twin-Incoherent Self-Expressive Locality-Adaptive Latent Dictionary Pair Learning for Classification56
- Robust Subspace Clustering With Compressed Data54
The proposed method, row space pursuit (RSP), recovers the authentic row space that gives correct clustering results under certain conditions, and is distinctly better than the competing methods, in terms of both clustering accuracy and computational efficiency.
- 53
- Motion control of the wedge prisms in Risley-prism-based beam steering system for precise target tracking52
It is found that the maximum required rotational velocities of the two prisms for target tracking are nearly the same and are dependent on the altitude angle and the central limited region is almost independent of the prism parameters.
- Fast and Robust Subspace Clustering Using Random Projections51
The proposed method, row space pursuit (RSP), recovers the authentic row space that gives correct clustering results under certain conditions, and is distinctly better than the competing methods, in terms of both clustering accuracy and computational efficiency.
- 49
- 46
- Multi-output regression on the output manifold46
This paper considers the output space as a Riemannian submanifold to incorporate its geometric structure into the regression process and proposes a novel mechanism, called locally linear transformation (LLT), to define the loss functions on the output manifold.
- 46
- A New Theory for Matrix Completion43
This paper introduces a new hypothesis called Isomeric condition, which is provably weaker than the assumption of uniform sampling and arguably holds even when the missing data is placed irregularly, and proves a series of theorems for missing data recovery and matrix completion.
- 39
- Flexible Auto-Weighted Local-Coordinate Concept Factorization: A Robust Framework for Unsupervised Clustering39
A novel unsupervised Robust Flexible Auto-weighted Local-coordinate Concept Factorization framework is proposed for clustering high-dimensional data, which improves the representations by enhancing the robustness of CF to noise and errors, providing a flexible constraint on the reconstruction error and optimizing the locality jointly.
- 37
- Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space37
The robust dictionary learning (DL) is investigated to discover the hybrid salient low-rank and sparse representation in a factorized compressed space and a discriminative J-RFDL model is presented, which can improve the discriminating abilities of learnt coefficients by minimizing the classification error jointly.
- On fusing the latent deep CNN feature for image classification35
This paper proposes a method to fuse the latent features extracted from the middle layers in a CNN to train a more robust classifier, and uses the late fusion strategy to combine the prediction of these classifiers.
- 35
- 34
- Hypergraph Embedding for Spatial-Spectral Joint Feature Extraction in Hyperspectral Images34
A novel algorithm for spatial-spectral feature extraction based on hypergraph embedding to ease the curse of the dimensionality problem and highlight the importance of spatial–spectral joint features embedding for the accurate classification of HSI data.
- 34
- Recovery of Future Data via Convolution Nuclear Norm Minimization33
This work converts TSF into a more inclusive problem called tensor completion with arbitrary sampling (TCAS), and proves that CNNM succeeds in solving TCAS as long as a sampling condition—which depends on the convolution rank of the target tensor—is obeyed.
- Listen, Think and Listen Again: Capturing Top-down Auditory Attention for Speaker-independent Speech Separation32
A novel neural network based structure motivated by the top-down attention behavior of human when facing complicated acoustical scene is presented, which gets rid of the limitation that the number of channels must be given or the high computation complexity for label permutation problem.
- Pointing error analysis of Risley-prism-based beam steering system32
Based on the vector form Snell's law, ray tracing is performed to quantify the pointing errors of Risley-prism-based beam steering systems, induced by component errors, prism orientation errors, and assembly errors.
- 32
- Robust auto-weighted projective low-rank and sparse recovery for visual representation31
This paper proposes an inductive Robust Auto-weighted Low-Rank and Sparse Representation framework by joint feature embedding for the salient feature extraction of high-dimensional data and demonstrates the effectiveness of the model for data representation and classification.
- 31
- 30
- 30
- 29
- Moving object detection via segmentation and saliency constrained RPCA29
This work proposes in this work a novel method termed Segmentation and Saliency constrained RPCA (SSC-RPCA), the solution of which is regularized by both segmentation and saliency constraints.
- MSSD: multi-scale self-distillation for object detection28
This work applies the idea of self-distillation to the object detection task and proposes a multi-scale self-distillation approach, where it is argued that knowledge distillation of the information contained in feature maps at different scales can help the model better detect small targets.
- Laser scanning by rotating polarization gratings27
It is shown that the scanning function in 1D scanning evolves from a sinusoidal to triangular scan and the duty cycle increases rapidly as the ratio of grating period to wavelength is reduced toward 2.
- Multi-level feature fusion pyramid network for object detection26
A Multi-Level Feature Fusion Pyramid Network (MLFFPN) that can fuse features with different receptive fields so as to produce reliable object representations robust against scale variation is designed.
- Matrix Completion with Deterministic Sampling: Theories and Methods26
This paper proposes two conditions, isomeric condition and relative well-conditionedness, for guaranteeing an arbitrary matrix to be recoverable from a sampling of the matrix entries and proves a collection of theorems for missing data recovery as well as convex/nonconvex matrix completion.
- Modeling Attention and Memory for Auditory Selection in a Cocktail Party Environment26
A unified Auditory Selection framework with Attention and Memory (dubbed ASAM) is proposed, which accumulates the prior knowledge into a life-long memory during the training phase, meanwhile a speech perceptor is trained to extract the temporal acoustic feature and update the memory online when a salient speech is given.
- Fine-grained action recognition using multi-view attentions24
This paper proposes a novel multi-view attention mechanism, named channel–spatial–temporal attention (CSTA) block, to guide the network to pay more attention to the clues useful for fine-grained action recognition, outperforming many state-of-the-art methods.
- 24
- 22
- 22
- Time Series Forecasting via Learning Convolutionally Low-Rank Models20
This paper integrates a learnable, orthonormal transformation into CNNM, with the purpose for converting the series of involute structures into regular signals of convolutionally low-rank, and proves that the resultant model, termed Learning-Based CNNM (LbCNNM), strictly succeeds in identifying the future part of a series.
- 20
- 18
- Learning Structured Twin-Incoherent Twin-Projective Latent Dictionary Pairs for Classification17
A novel framework called Twin-Projective Latent Flexible DPL (TP-DPL) is proposed, which minimizes the twin-incoherence constrained flexibly-relaxed reconstruction error to avoid the possible over-fitting issue and produce accurate reconstruction.
- Robust Unsupervised Flexible Auto-weighted Local-coordinate Concept Factorization for Image Clustering17
This work investigates the high-dimensional data clustering problem by proposing a novel and unsupervised representation learning model called Robust Flexible Auto-weighted Local-coordinate Concept Factorization (RFA-LCF), which integrates the robust flexible CF, robust sparse local-coordinated coding and the adaptive reconstruction weighting learning into a unified model.
- Deep Latent Low-Rank Fusion Network for Progressive Subspace Discovery16
A new and effective strategy to extend the sin-gle-layer latent low-rank models into multi-ple-layers, and a new and progressive Deep Latent Low-Rank Fusion Network (DLRF-Net) to uncover deep features and struc-tures embedded in input data are presented.
- Real-Time Traffic Sign Detection Based on Weighted Attention and Model Refinement15
This work improves the current state-of-the-art detection algorithm YOLOv7 as a basis and proposes a lightweight convolution-based spatial pyramidal pooling fusion module (DW-SPPF) and replaces the conventional convolution of the detection head with a lightweight depth-separable convolution to reduce the complexity of the model while maintaining detection accuracy.
- ChipSong: A Controllable Lyric Generation System for Chinese Popular Song15
ChipSong is proposed, which is an assisted lyric generation system built based on a Transformer-based autoregressive language model architecture, and generates controlled lyric paragraphs fit for musical short-video display purpose, by designing a novel Begin-Internal-End word-granularity embedding sequence.
- Beam steering limitation of a Risley prism system due to total internal reflection15
It is shown that the ray deviation power reaches its extreme value when the opening angles increase to a certain limit value for a given prism material, which can afford guidance for prism material and geometry choices in the design of a wide-angle Risley-prism-based beam steering system.
- Robust Low-rank subspace segmentation with finite mixture noise14
This paper focuses on the complex noise contaminated subspace clustering problem by using finite mixture of exponential power (MoEP) distributions, and introduces a novel prior on the singular values of representation matrix, which leads to a novel penalty in the nonconvex and nonsmooth optimization.
- 12
- Matrix recovery with implicitly low-rank data12
A novel method is proposed for matrix recovery that pursues the low-rank structure of the target matrix in an implicit feature space by making use of the specifics of an accelerated proximal gradient based optimization algorithm.
- Multi-component group sparse RPCA model for motion object detection under complex dynamic background12
A multi-component group sparse RPCA model is proposed to cope with all the difficulties of robust PCA and an incoherence term is introduced to enhance the separability of sparse foreground motion from dynamic background component.
- Latent Low-Rank Representation12
This chapter proposes to construct the dictionary by using both observed and unobserved, hidden data, and shows that the effects of the hidden data can be approximately recovered by solving a nuclear norm minimization problem, which is convex and can be solved efficiently.
- Semi-supervised vehicle classification via fusing affinity matrices11
A semi-supervised learning model is used to classify vehicle types, which only needs a small number of pre-labeled data and propagates these labels to the rest data at hand and results support the effectiveness of the method.
- 11
- Artificial intelligence-assisted remote sensing observation, understanding, and decision10
This commentary examines how AI enhances visual observation, facilitates semantic transition, and empowers intelligent decision-making, which provide support for the paradigm shift from data acquisition to cognitive services.
- Similarity-Adaptive Latent Low-Rank Representation for Robust Data Representation10
A novel Similarity-Adaptive Latent Low-Rank Representation model that incorporates a reconstructive error minimization term over coefficients and salient features, which can clearly preserve the neighborhood information of salient features adaptively and make the coefficients coding process very efficient.
- Decoupled spatio-temporal grouping transformer for skeleton-based action recognition9
A novel decoupled spatio-temporal grouping Transformer (DSTGFormer) model is presented, which outperforms the state-of-the-art methods on two large-scale datasets in terms of both recognition accuracy and computational efficiency.
- Multilayer Collaborative Low-Rank Coding Network for Robust Deep Subspace Discovery9
A new Multilayer Collabora-tive Low-Rank Representation Network model termed DeepLRR is proposed to discover deep features and deep subspaces to recover the sparse errors in layers accurately and outperforms other related models for subspace discovery and clustering.
- 9
- 3D Tensor Auto-encoder with Application to Video Compression8
This article introduces a network structure called 3D Tensor Auto-Encoder (3DTAE), in which a video is represented as a vector, and the weights of each layer are represented by three small matrices, and thus the number of parameters in the network is just O(n1/3).
- A Latent Clothing Attribute Approach for Human Pose Estimation8
This paper proposes and explores a latent clothing attribute approach for HPE that models the clothing attributes as latent variables and thus requires no explicit labeling for the Clothing attributes.
- 7
- 7
- Learning Additive Exponential Family Graphical Models via $\ell_{2,1}$-norm Regularized M-Estimation7
Two -norm regularized maximum likelihood estimators to learn the model parameters from i.i.d. samples are proposed and it is shown that under mild conditions the extra flexibility gained by the additive exponential family models comes at almost no cost of statistical efficiency.
- Non-blind deblurring of structured images with geometric deformation7
This paper shows how to enhance non-blind deconvolution by recovering and undoing the deformation while deconvolving a given blurred image and proves that almost all popular regularizers that have been proposed for image deblurring such as total variation and its variants can be improved.
- 7
- A real-time semi-dense depth-guided depth completion network6
This work proposes a simple yet effective algorithm that aims to combine the strengths of both the traditional image processing techniques and the prevalent deep learning methods to predict dense depth maps from given depth maps of sparse.
- 6
- 6
- Correction to: On fusing the latent deep CNN feature for image classification6
The original version of this articleUnfortunately, the spelling of Guangcan Liu’s name was incorrect.
- 6
- 5
- SSCNet: learning-based subspace clustering4
This work redefine and generalize the optimization procedure of the linearized alternating direction method of multipliers (ADMM), framing it as a multi-block deep neural network, where each block corresponds to a linearized ADMM iteration step, to address the SSC problem.
- Collaborative Local-Global Learning for Temporal Action Proposal4
CLGNet is an integration of Temporal Convolution Network and Bidirectional Long Short-Term Memory, in which Temporal Convolution Network is responsible for local dependencies while Bidirectional Long Short-Term Memory takes charge of handling the global dependencies.
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- Test-Time Scaling with Reflective Generative Model3
This work introduces the first reflective generative model MetaStone-S1, which obtains OpenAI o3-mini's performance via the new Reflective Generative Form, and provides a self-supervised process reward model which can directly learn the high-quality reasoning trajectory selection from the outcome reward.
- 3
- DLRF-Net: A Progressive Deep Latent Low-Rank Fusion Network for Hierarchical Subspace Discovery3
This article proposes a new progressive Deep Latent Low-Rank Fusion Network (DLRF-Net) to uncover deep features and the clustering structures embedded in latent subspaces and presents a novel strategy that can extend existing single-layer latent low-rank models into multiple layers.
- 3
- 3
- 3
- 2
- Presswork defect inspection using only defect-free high-resolution images2
This paper proposes an end-to-end cropping method to balance the number of samples which cropped from different patterns in the defect-free high-resolution images, and thus, can solve the quantity imbalance between different patterns (image level).
- 2
- Tensor LISTA: Differentiable sparse representation learning for multi-dimensional tensor2
Benefiting from tensor representation and differentiable programming, TLISTA achieves rapid inference speed and acquires more valuable representation for the data primarily organized in tensor form.
- 2
- 2
- Matrix Completion with Nonuniform Sampling: Theories and Methods.2
This work introduces a new hypothesis called isomeric condition, which is provably weaker than the assumption of uniform sampling, and proves that the exact solutions that identify the target matrix are included as critical points by the commonly used bilinear programs.
- Spatial–spectral locality constrained elastic net hypergraph for hyperspectral image clustering2
A new spatial–spectral locality constrained elastic net hypergraph learning model is proposed for hyperspectral image clustering and can achieve a superior clustering performance when compared to state-of-the-art methods.
- Scalable Low-Rank Representation2
This chapter addresses the problem of solving nuclear norm regularized optimization problems (NNROPs), which contain a category of problems including LRR, and revisit the classic mechanism of low-rank matrix factorization, based on which an active subspace algorithm is presented for efficiently solving NNROPs.
- 2
- Home Real-Time Remote Monitoring System Based on Embedded Linux2
This system relies on the mature embedded technology and achieves the embedded Linux-based real-time remote monitoring system on the Mini2440 development board in accordance with the above design.
- 2
- Parameter Optimization of RBF Networks Based on Evolutionary Model2
A computation model of a population-based algorithm-generator is proposed to estimate the parameters of the RBF networks and the advantage of the model is the functional decomposition.
- Research for the orientation detection system using L-shape reticle2
The research result shows that the orientation detection system using L-shape reticle has a good effect on space-filtering, the signals that it sends out are easy to be processed by computer, its self-adapting tracking gate has a strong anti-interference ability, and the whole system's searching and tracking performances are quite high.
- 1
- 1
- Auto-Focus Contrastive Learning for Image Manipulation Detection1
An Auto-Focus Contrastive Learning (AF-CL) network for image manipulation detection contains two main ideas, i.e., multi-scale view generation (MSVG) and trace relation modeling (TRM), which aims to generate a pair of views, each of which contains the manipulated region and its surroundings at a different scale.
- Advancing Iterative Quantization Hashing Using Isotropic Prior1
This paper proposes a new hashing method, termed Isotropic Iterative Quantization (IITQ), that extends the formulation of ITQ by incorporating properly the isotropic prior proposed by [3], and devise a proximal method that can solve problem in a practical fashion.
- GPS Navigation Information Processing and Display Based on VC++1
On the basis of the positioning principle of GPS, the GPS navigation information processing method is proposed which has combined the PC and the receiver serial communication.
- Design of Layered Aerospace Time Synchronization Architecture and Time Synchronization Links Budget Comparison1
The results show that GEO-GEO link can achieve to more than 3Gb/s data transmission rate on condition that the transmitting power is 0.5W, and the antenna diameter is only 20cm, which serves as reference for the establishment of layered aerospace time synchronization architecture and improvement of time synchronization precision.
- 1
- Evolving Ridgelet Neural Network and Its Application1
The ridgelet netowrk model estimated by the proposed optimazation algorithm is used to predict the famous Mackey-Glass time series.
- 1
- –
- –
- –
- Class activation map replay and minimum entropy sampling-based multilabel class-incremental learning–
- –
- Towards stable training of parallel continual learning–
Stable Parallel Continual Learning (SPCL) is introduced, a novel approach that enhances the training stability of PCL for both forward and backward propagation and outperforms state-of-the-art methjods and achieve better training stability.
- Parallel Connected LSTM for Matrix Sequence Prediction with Elusive Correlations–
This article proposes a novel architecture called Parallel Connected LSTM (PcLSTM), which integrates two new mechanisms, Multi-channel Linearized Connection (McLC) and Adaptive Parallel Unit (APU), into the framework of L STM, and is able to handle well both the elusive correlations within each timestamp and the temporal dependencies across different timestamps.
- –
- Pedestrian Detection with a Directly-Cascaded Deconvolution-Convolution Structure–
A novel neural network structure which can be flexibly combined with powerful object detection systems for boosting pedestrian detection and incorporates the features learned in different layers in a novel and unified fashion is proposed.
- –
- –
- –
- –
- –
- –
- High Precision Voltage Adjustable DC-DC Power Supply Based on Function Control–
The experiment shows that the DC-DC power supply improves the control precision for voltage output with using the control function.
- Anti infrared interference information processing method of the infrared angular instrument–
This paper discussed the anti-infrared interference information processing method of the infrared angular instrument, and proposed the signal data overflow handling algorithm to deal with the possible missile target missing caused by the infrared interference, and the infrared signal flicker because of therared attenuation in atmosphere.
- Using C++ to Drive a Simulation Elevator Control System–
This thesis generally introduced the method of using C++ programming lanuage to simulate an elevator system,ounded the algorithm and data structure which was used in the method, and briefly introduced some instructions according to the software engineering.
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.