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
- Songde MaSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works352 from public data
- Stacked Sparse Autoencoder (SSAE) for Nuclei Detection on Breast Cancer Histopathology Images851
A Stacked Sparse Autoencoder, an instance of a deep learning strategy, is presented for efficient nuclei detection on high-resolution histopathological images of breast cancer and out-performed nine other state of the art nuclear detection strategies.
- Fast Visual Tracking via Dense Spatio-temporal Context Learning705
A novel explicit scale adaptation scheme is proposed, able to deal with target scale variations efficiently and effectively, and the Fast Fourier Transform is adopted for fast learning and detection in this work, which only needs 4 FFT operations.
- Classification of Hyperspectral and LiDAR Data Using Coupled CNNs463
An efficient and effective framework to fuse hyperspectral and light detection and ranging (LiDAR) data using two coupled convolutional neural networks (CNNs) designed to learn spectral–spatial features from hyperspectrals and elevation information from LiDAR data is proposed.
- 394
- 374
- 354
- 347
- 343
- Hyperspectral Image Classification With Attention-Aided CNNs340
An attention-aided CNN model based on the traditional CNN model that incorporates attention modules to aid networks that focus on more discriminative channels or positions for spectral and spatial classifications of hyperspectral images is proposed.
- Robust Lightweight Facial Expression Recognition Network with Label Distribution Training326
This paper presents an efficiently robust facial expression recognition (FER) network, named EfficientFace, which holds much fewer parameters but more robust to the FER in the wild, and introduces a simple but efficient label distribution learning (LDL) method as a novel training strategy.
- 305
- 280
- 251
- 249
- Former-DFER: Dynamic Facial Expression Recognition Transformer246
The heatmaps of the leaned facial features demonstrate that the proposed Former-DFER is capable of handling the issues such as occlusion, non-frontal pose, and head motion, and the visualization of the feature distribution shows that the propose method can learn more discriminative facial features.
- 216
- 173
- 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.
- 157
- Ensemble learning for independent component analysis149
An enhanced ICA algorithm by ensemble learning approach, named as random independent subspace (RIS), to deal with the two problems of unstable ICA classifier and small sample size problem is proposed.
- 144
- 144
- 141
- 131
- A Variational Approach to Simultaneous Image Segmentation and Bias Correction119
An efficient iterative algorithm is proposed for energy minimization, via which the image segmentation and bias field correction are simultaneously achieved and the smoothness of the obtained optimal bias field is ensured by the normalized convolutions without extra cost.
- Adaptive Graph Convolutional Network With Attention Graph Clustering for Co-Saliency Detection110
This work presents a novel adaptive graph convolutional network with attention graph clustering (GCAGC), and proposes a unified framework with encoder-decoder structure to jointly train and optimize the graph convolutionsal network, attention graph cluster, and co-saliency detection decoder in an end-to-end manner.
- 106
- 104
- Boosting encoded dynamic features for facial expression recognition104
This paper designs the dynamic haar-like features to represent the temporal variations of facial appearance and further encode the dynamic features into the binary pattern features, which are useful to construct weak classifiers for boosting learning.
- 102
- 101
- 99
- 94
- 94
- 94
- Self-Paced Multi-Task Learning91
This paper proposes a novel multi-task learning framework that attempts to learn the tasks by simultaneously taking into consideration the complexities of both tasks and instances per task, by presenting a new task-oriented regularizer that can jointly prioritize tasks and instance.
- Temporal spectral residual91
A new method Temporal Spectral Residual on video slices along X-T and Y-T planes, which can automatically separate foreground motion objects from backgrounds, also with the help of threshold selection and voting schemes is proposed.
- Adaptive Multi-View and Temporal Fusing Transformer for 3D Human Pose Estimation89
A unified framework to adaptively handle varying view numbers and video length without camera calibration in 3D Human Pose Estimation (HPE) is proposed, which obtains competitive results and generalizes well to dynamic capture with an arbitrary number of unseen views.
- FRNet: Frustum-Range Networks for Scalable LiDAR Segmentation88
FRNet is a simple yet powerful method aimed at restoring the contextual information of range image pixels using corresponding frustum LiDAR points using corresponding frustum LiDAR points, and operates 5 times faster than state-of-the-art approaches.
- VisDrone-SOT2018: The Vision Meets Drone Single-Object Tracking Challenge Results87
The evaluation protocol of the VisDrone-SOT2018 challenge is presented and the results of a comparison of 22 trackers on the benchmark dataset are presented, which are publicly available on the challenge website.
- 87
- 85
- 85
- 84
- 81
- 78
- 76
- 76
- 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
- 71
- A One-Layer Recurrent Neural Network for Real-Time Portfolio Optimization With Probability Criterion71
- 70
- 67
- Facial expression recognition using encoded dynamic features65
A novel framework for video-based facial expression recognition, which can handle the data with various time resolution including a single frame, is proposed, which uses the haar-like features to represent facial appearance and performs K-Means clustering on the facial appearance features to explore the intrinsic temporal patterns of each expression.
- 63
- 61
- 60
- Learning Memory Augmented Cascading Network for Compressed Sensing of Images60
A cascading network for compressed sensing of images with progressive reconstruction by decomposing the complex reconstruction mapping into the cascade of incremental detail reconstruction (IDR) modules and measurement residual updating (MRU) modules.
- 60
- Hypergraph with sampling for image retrieval59
A new transductive learning framework for image retrieval is proposed, in which images are taken as vertices in a weighted hypergraph and the task of image search is formulated as the problem of hypergraph ranking.
- CNN-based near-real-time precipitation estimation from Fengyun-2 satellite over Xinjiang, China58
An effective and efficient model based on convolutional neural network (CNN) to estimate precipitation over Xinjiang, the driest region in China, from Fengyun-2 satellite data is proposed and can achieve near-real-time results, which is valuable for precipitation nowcasting applications.
- M3 CSR: Multi-view, multi-scale and multi-component cascade shape regression58
A multi-view, multi-scale and multi-component cascade shape regression (M3CSR) model for robust face alignment is presented and a component-based shape refinement process is developed to further improve the performance of face alignment.
- 56
- 55
- 54
- 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.
- 52
- 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.
- Immune network-based swarm intelligence and its application to unmanned aerial vehicle (UAV) swarm coordination51
A multi-agent cooperative operation strategy for the immune system based on the immune network theory is proposed and each agent is designed to select its favorable strategy according to distributed sensing through local communication to achieve successful swarm coordination.
- Robust visual tracking via patch based kernel correlation filters with adaptive multiple feature ensemble50
A patch based tracker which adaptively integrates the kernel correlation filters with multiple effective features to learn the correlations between the target and the background, the candidate patches and template ones, which boosts the overall performance.
- 49
- 47
- 46
- 46
- 46
- 45
- 45
- 44
- Deep Learning for Histopathological Image Analysis: Towards Computerized Diagnosis on Cancers44
This chapter presents deep learning based approaches for two challenged tasks in histological image analysis: Automated nuclear atypia scoring (NAS) on breast histopathology and Epithelial and stromal tissues discrimination.
- Deep Object Co-Segmentation via Spatial-Semantic Network Modulation43
A spatial and semantic modulated deep network framework for object co-segmentation that captures the correlations of image feature descriptors via unsupervised learning and proposes a hierarchical second-order pooling module to transform the image features for classification use.
- 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.
- 42
- A Self-Paced Regularization Framework for Multilabel Learning42
This work proposes a new multilabel learning formulation by introducing a self-paced function as a regularizer, so as to simultaneously prioritize label learning tasks and instances in each iteration, and provides a general way to find the desiredSelf-paced functions.
- 41
- 39
- 39
- 38
- 36
- 36
- 35
- Defenses in Adversarial Machine Learning: A Survey35
This survey aims to build a systematic review of all existing defense paradigms from a unified perspective, and factorize a complete machine learning system into five stages, including pre-training, training, post-training, deployment, and inference stages, respectively.
- 34
- 34
- ProxyBNN: Learning Binarized Neural Networks via Proxy Matrices33
By introducing an appropriate proxy matrix, this paper reduces the weights quantization error while circumventing explicit binary regularizations on the full-precision auxiliary variables and is the first to empirically identify and study the effectiveness of learning both basis and coefficients to construct the pre-binarization weights.
- Visual tracking via Boolean map representations33
A simple yet effective Boolean map based representation that exploits connectivity cues for visual tracking that facilitates achieving favorable performance in terms of accuracy and robustness against the state-of-the-art tracking methods on the OTB50 and VOT2016 benchmark datasets.
- 33
- Robust object tracking by online Fisher discrimination boosting feature selection33
This paper proposes an online Fisher discrimination boosting feature selection mechanism, which selects features that reduce the with-in scatter while enlarging the between-class scatter, thereby enhancing the discriminative capability between the target and background.
- 33
- 32
- 32
- Meta-learning with Network Pruning31
A network pruning based meta-learning approach for overfitting reduction via explicitly controlling the capacity of network is proposed that not only effectively alleviates meta-overfitting but also in many cases improves the overall generalization performance when applied to few-shot classification tasks.
- 31
- 30
- 29
- 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.
- 29
- Unsupervised Spatial–Spectral Network Learning for Hyperspectral Compressive Snapshot Reconstruction28
An unsupervised spatial–spectral network to reconstruct HSIs only from the compressive snapshot measurement is proposed, which acts as a conditional generative model conditioned on the snapshot measurement, and it exploits the spatial—spectral attention module to capture the joint spatial-spectral correlation of HSIs.
- 27
- 26
- 26
- 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.
- Single image super-resolution with enhanced Laplacian pyramid network via conditional generative adversarial learning26
An enhanced Laplacian pyramid network (ELapN) is designed as the generator model of CGAN for SISR to better reconstruct image with a large-scale factor and demonstrates superiority of the proposed method over state-of-the-art methods in terms of PSNR, SSIM and visual effect.
- Pan-GGF: A probabilistic method for pan-sharpening with gradient domain guided image filtering26
The proposed novel probabilistic pan-sharpening method with gradient domain guided image filtering (named Pan-GGF) shows better spatial and spectral fusion than other fusion methods, such as projection substitution, detail injection and weighted combination models.
- 26
- Robust facial landmark tracking via cascade regression26
A robust facial landmark tracking system via cascade regression, which is able to deal well with some challenges emerging in the sequential images, and achieves competing results on the LFPW, Helen, 300-W and 300-VW datasets.
- 26
- 25
- 23
- Statistical Downscaling of Temperature Distributions in Southwest China by Using Terrain-Guided Attention Network23
In general, among various deep learning approaches, TGAN achieves better downscaling results for 2 m air temperature reconstruction and provides a practical method and guidance for the back-calculation of high-resolution historical meteorological grid data.
- Adaptive Compressive Tracking via Online Vector Boosting Feature Selection23
An adaptive CT approach is proposed, which selects the most discriminative features to design an effective appearance model and a multiple scale adaptation mechanism is explored to estimate object size, which helps to relieve interference from background information.
- Patch-based active learning (PTAL) for spectral-spatial classification on hyperspectral data23
Experimental results show better performance of the proposed PTAL methods on classification accuracy and computational time on three different hyperspectral data sets as compared with PXAL methods.
- 22
- 22
- 21
- Image super‐resolution using conditional generative adversarial network21
An SISR approach based on conditional GAN (SRCGAN), which includes a generator network that generates super-resolution (SR) images and a discriminator network that is trained to distinguish the real images from the SR images, facilitating training a more stable generator model than GAN without this guidance.
- 20
- Face image retrieval based on shape and texture feature fusion20
An efficient method is proposed that uses accurate facial landmark locations as shape features and shape priors to provide discriminative texture features for convolutional neural networks to solve face image retrieval problems.
- Neural Information Processing20
The results reveal that the SVR controller together with SVR model leads to good tracking performance with small modeling, transient state and steady state errors.
- 19
- 19
- Monitoring activity of taking medicine by incorporating RFID and video analysis19
A new framework to monitor medication intake for elderly individuals by incorporating a video camera and radio frequency identification (RFID) sensors is presented, demonstrating 100 % detection accuracy for identifying medicine bottles and promising results for monitoring activity of taking medicine.
- 18
- Identifying Regional Cardiac Abnormalities from Myocardial Strains Using Spatio-temporal Tensor Analysis18
A novel tensor-based classification framework is developed that better conserves the spatio-temporal structure of the myocardial deformation pattern than conventional vector-based algorithms and keeps more of the information of the original feature space, so that abnormal tensors in the subspace can be back-projected to reveal the regional cardiac abnormality in a more physically meaningful way.
- 18
- An Empirical Study on Data Augmentation for Pixelwise Satellite Image Time-Series Classification and Cross-Year Adaptation17
This study provides an empirical study of 11 alternative augmentation techniques for pixelwise satellite time series, including noise injection, scaling, mixup, weighted dynamic time warping barycentric averaging, temporal dropout, window slicing, temporal shift, time warping, interpolation resampling, amplitude jittering, and phase jittering.
- Learning residual refinement network with semantic context representation for real-time saliency object detection17
A residual refinement network with semantic context features for SOD is proposed, which can not only gradually recover the spatial details in each feature map from top to down, but also enhance the features at all scales with high-level semantic context information.
- 17
- 17
- Robust Tracking via Convolutional Networks without Learning.17
It is presented that, even without learning, simple convolutional networks can be powerful enough to develop a robust representation for visual tracking, and it is shown that this representation can robustly adapt to target appearance variations.
- 17
- Dual Temporal Memory Network for Efficient Video Object Segmentation16
An end-to-end network which stores short- and long-term video sequence information preceding the current frame as the temporal memories to address the temporal modeling in VOS is presented.
- Integrating Convolutional Neural Network and Gated Recurrent Unit for Hyperspectral Image Spectral-Spatial Classification16
Experimental results show that the proposed method can achieve better performance than comparison methods, and is compared with several state-of-the-art deep learning methods on two real HSIs.
- Cascaded Regional Spatio-Temporal Feature-Routing Networks for Video Object Detection16
This paper incorporates the correlation filter tracking on the convolutional feature maps to explore an efficient and effective spatiotemporal region proposal generation method, and proposes a router function which selects feature maps according to the scale of proposals.
- 16
- A Framework for the Recognition of Nonmanual Markers in Segmented Sequences of American Sign Language16
This work proposes a novel framework for robust tracking and analysis of non-manual behaviours, with an application to sign language recognition, and applies it to segmented recognition of five classes of grammatical constructions conveyed through facial expressions and head gestures.
- 16
- 15
- 15
- 15
- 15
- Dynamic soft encoded patterns for facial event analysis15
This paper develops similarity features to describe complicated variations of facial appearance, which take similarities between a haar-like feature in a given image and the corresponding ones in reference images as feature vector, and maps them into dynamic soft encoded pattern (DSEP) for facial event analysis.
- 15
- 14
- 14
- 14
- Bidirectionally Learning Dense Spatio-temporal Feature Propagation Network for Unsupervised Video Object Segmentation14
This paper presents a novel Dense Bidirectional Spatio-temporal feature propagation Network (DBSNet) to fully integrate the forward and the backward propagations across all frames and achieves favorable performance against state-of-the-art methods in terms of all evaluation metrics.
- 14
- 14
- FaceHunter: A multi-task convolutional neural network based face detector14
The proposed FaceHunter is evaluated on the AFW dataset, FDDB dataset and Pascal Faces respectively, and extensive experiments demonstrate its powerful performance against several state-of-the-art detectors.
- Decentralized Robust Subspace Clustering14
A large scale distributed framework for the computation of SSC via an alternating direction method of multiplier (ADMM) algorithm that solves SSC in column blocks and only involves parallel multivariate Lasso regression subproblems and sample-wise operations.
- Special Section on Object and Event Classification in Large-Scale Video Collections14
The nine papers in this special section on object and event classification in large-scale video collections can be categorized into four themes: video indexing, concept detection, video summarization, and event recognition.
- 14
- 14
- 14
- A Unified Object Counting Network With Object Occupation Prior13
This paper builds the first evolving object counting dataset and proposes a unified object counting network that can efficiently and effectively adapt to new classes while maintaining good performance on already-seen data without large-scale retraining.
- Intellectual property protection for deep semantic segmentation models13
This work designs a new hybrid intellectual property protection framework by combining the trigger-set based and passport based watermarking simultaneously, and presents the first attempt at protecting deep semantic segmentation models from potential infringements.
- 13
- 13
- 13
- Lesion-Specific Coronary Artery Calcium Quantification for Predicting Cardiac Event with Multiple Instance Support Vector Machines13
A lesion-specific CAC quantification framework to improve CAC's near-term predictive value in intermediate to high-risk populations with a novel multiple instance support vector machines (MISVM) approach is proposed.
- Attacks in Adversarial Machine Learning: A Systematic Survey from the Lifecycle Perspective12
It is demonstrated that isolated defenses are insufficient against coordinated multi-stage attacks, and the research community must prioritize holistic defense strategies incorporating lifecycle-aware monitoring, adaptive hardening techniques, and unified threat models.
- Geometric Understanding of Discriminability and Transferability for Visual Domain Adaptation12
A geometry-oriented model is proposed for enhancing the transferability and discriminability via nuclear norm optimization and extensive experiment results validate the effectiveness of the proposed model in empirical applications, and verify that the geometric abilities can be sufficiently learned in the derived feasible range.
- 12
- Learning lightweight Multi-Scale Feedback Residual network for single image super-resolution12
A lightweight feedback-based recurrent neural network (FRNN) tailored to SISR is designed, which achieves favorable performance on five benchmark datasets compared to the state-of-the-art methods in terms of PSNR and SSIM.
- Max-Margin-Based Discriminative Feature Learning12
A new max-margin-based discriminative feature learning method is proposed, which aims at learning a low-dimensional feature representation so as to maximize the global margin of the data and make the samples from the same class as close as possible.
- Ordinal regularized manifold feature extraction for image ranking12
A novel feature extraction algorithm specifically designed for learning to rank in image ranking not only targets at preserving the local manifold structure of data, but also keeps the ordinal information among different data blocks in the low-dimensional subspace, where a ranking model can be learned effectively and efficiently.
- Riemannian manifold-based disentangled representation learning for multi-site functional connectivity analysis11
A novel Riemannian Manifold-based Disentangled Representation Learning (RM-DRL) framework which is capable of learning invariant representations from fMRI data across multiple sites for brain disorder diagnosis, and is evaluated on the public multi-site ABIDE dataset.
- 11
- 11
- 11
- 11
- 11
- 11
- Inferring occluded features for fast object detection11
The experimental results demonstrate the effectiveness of the proposed method both in improving the classification performance and achieving a fairly high processing speed.
- Low rank driven robust facial landmark regression11
Extensive results on several wild benchmarks such as COFW, LFPW and Helen demonstrate that the proposed method is robust to facial occlusions, pose variations and exaggerated facial expressions.
- 11
- 10
- 10
- 10
- 10
- 10
- Convolutional Neural Networks for Clothes Categories10
This paper designs convolutional neural networks(CNNs) architecture and adaptively learns the feature representation of clothes from the authors' constructed dataset and demonstrates the superiority of the CNNs to the other algorithms.
- A new extension of kernel feature and its application for visual recognition10
The kernel feature vector of an image is extended to a kernel feature matrix for visual recognition using multiple similarities to measure two images, which is more accordant to human vision.
- DuPt: Rehearsal-based continual learning with dual prompts9
This work proposes an input-aware prompt, an input-level cue that utilizes an input prior to querying for valid cue information that can act as a consistency regularization to avoid feature conflicts caused by the differences between network features.
- A high-deck coach evacuation model framework: Behavioural modelling, numerical analyses and insights9
- Single-atom nanozymes with intelligent response to pathological microenvironments for bacterially infected wound healing9
A full-thickness skin wound-infected model demonstrates that MN4 single-atom nanozymes accelerate wound repair and skin regeneration by suppressing the expression of tumor necrosis factor-alpha, promoting angiogenesis, and enhancing collagen deposition.
- U4D: Uncertainty-Aware 4D World Modeling from LiDAR Sequences9
U4D is presented, an uncertainty-aware framework for 4D LiDAR world modeling that produces geometrically faithful and temporally consistent LiDAR sequences, advancing the reliability of 4D world modeling for autonomous perception and simulation.
- 9
- 9
- 9
- Genome‑wide identification and characterization of miR396 family members and their target genes GRF in sorghum (Sorghum bicolor (L.) moench)9
The results showed that five Sbi-miR396 members, located on chromosomes 4, 6, and 10, were identified at the whole-genome level, and target gene prediction indicates that the growth-regulating factor family members (SbiGRF1/2/3/4/5/6/7/8/10) were target genes of SBi-mi-396d/e.
- Consistent connectome landscape mining for cross-site brain disease identification using functional MRI9
A novel Connectome Landscape Modeling (CLM) method that can mine cross-site consistent connectome landscape and extract data-driven representation of functional connectivity networks for brain disorder identification and an efficient algorithm based on the Alternating Direction Method of Multipliers (ADMM) to solve the proposed objective function are developed.
- 9
- Dual iterative hard thresholding9
A novel duality theory is developed that establishes a set of suficient and/or necessary conditions under which the original non-convex problem can be equivalently or approximately solved in a concave dual formulation and proposes the Dual IHT (DIHT) algorithm as a super-gradient ascent method to solve the non-smooth dual problem.
- MORES: Online Incremental Multiple-Output Regression for Data Streams.9
A novel online multiple-output regression method, called MORES, for streaming data, that can learn the structure of the regression coefficients to facilitate the model's continuous refinement and introduce a forgetting factor to set different weights on samples so as to track the data streams' evolving characteristics quickly from the latest samples.
- Video event description in scene context9
A framework for automated video event description is presented, which features fused with the context knowledge to provide accurate and reliable event description, designed to describe the event and recognize objects activities composed of four components.
- 9
- 8
- Spatiotemporal Enhanced Adversarial Network for Precipitation Nowcasting8
A spatiotemporal enhanced adversarial network (STEAN), a deep learning model for high-resolution precipitation nowcasting that leverages a specialized Halo Attention Spatiotemporal Long Short-Term Memory unit to model temporal variation information.
- 8
- CED-Net: contextual encoder–decoder network for 3D face reconstruction8
A contextual encoder–decoder network to regress UV position map for 3D face reconstruction, named CED-Net, which introduces contextual information both in the shape and feature level, which is served as the crucial information.
- Low-rank weighted co-saliency detection via efficient manifold ranking8
An effective co-Saliency detection approach that first exploits an efficient manifold ranking scheme to extract a set of co- saliency regions, and then renders rank constraint to the feature matrix of the extracted regions to achieve a high-quality co-saliency map is presented.
- Distributed Inexact Newton-type Pursuit for Non-convex Sparse Learning8
A sample distributed greedy pursuit method for non-convex sparse learning under cardinality constraint that alternates between local inexact sparse minimization of a Newton-type approximation and centralized global results aggregation.
- Parallel Sparse Subspace Clustering via Joint Sample and Parameter Blockwise Partition8
A memory-efficient parallel framework for computing SSC via an alternating direction method of multiplier (ADMM) algorithm that is substantially more efficient than its matrix counterpart used by SSC, without sacrificing performance in applications.
- 8
- Remote Sensing Object Counting With Online Knowledge Learning7
This work builds an end-to-end training framework that seamlessly integrates two distinct networks into a unified one and proposes a relation-in-relation distillation (RiRD) method that allows the student branch to effectively comprehend the evolution of the relationship of intralayer teacher features among different interlayer features.
- Cooperative Sentiment Agents for Multimodal Sentiment Analysis7
A new Multimodal Representation Learning (MRL) method for Multimodal Sentiment Analysis (MSA), which facilitates the adaptive interaction between modalities through Cooperative Sentiment Agents, named Co-SA, which excels at discovering diverse cross-modal features.
- 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.
- Fast compression and reconstruction of astronomical images based on compressed sensing7
A new real-time compression and fast reconstruction algorithm for astronomical images based on compressive sensing techniques that can efficiently reconstruct colorful astronomical images with high resolution and improve the applicability of compressed sensing.
- 7
- 7
- FauForensics: Boosting Audio-Visual Deepfake Detection With Facial Action Units6
FauForensics, a novel framework leveraging biologically invariant facial action units (FAUs), which are quantitative descriptors of facial muscle activity linked to emotion physiology, is proposed, which serve as forgery-resistant representations that reduce domain dependency while capturing subtle synthetic-content disruptions.
- Adaptive frequency collaboration for remote sensing change detection6
An adaptive frequency collaboration network (AFCN) is proposed to construct change features from a frequency-domain perspective to achieve frequency disentanglement and achieves state-of-the-art performance on three benchmark datasets.
- Temporally Efficient Gabor Transformer for Unsupervised Video Object Segmentation6
A Temporally efficient Gabor Transformer (TGFormer) for UVOS is presented, which makes the UVOS model be a fully Transformer architecture, meanwhile aggregating the information from space, Gabor and time domains, yielding a strong representation with rich structure details.
- 6
- Body parts relevance learning via expectation–maximization for human pose estimation6
A expectation–maximization (EM) mapping-based network to learn specific related body parts for human pose estimation, named EMposeNet, which aims at performing EM algorithm to learn the related body part, where the predicted keypoint heatmap is potentially more accurate and spatially more precise.
- 6
- 6
- Efficient k-support-norm regularized minimization via fully corrective frank-Wolfe method6
This paper reformulates the k-support-norm regularized formulation into an identical constrained formulation and proposes a fully corrective Frank-Wolfe algorithm to minimize the constrained model.
- Active Sample Learning and Feature Selection: A Unified Approach.6
This paper presents a unified framework to conduct active learning and feature selection simultaneously in a fully unsupervised setting by solving the NP-hardness of the raw problem efficiently by an iterative algorithm, and proves its global convergence.
- 6
- 6
- 6
- 6
- A Unified Framework of Subspace and Distance Metric Learning for Face Recognition6
This paper proposes a unified scheme of subspace and distance metric learning under the Bayesian framework for face recognition, and formulate it by minimizing the Bhattachyrra coefficient between two distributions.
- 6
- Joint Intensity and Spatio-Temporal Representation Learning for Extreme Precipitation Nowcasting5
An Intensity Trend Module (ITM) is designed to explicitly learn and model the intensity features of extreme precipitation events, and a Spatio-Temporal Coherence Module (STCM), which exploits voxel flow to capture spatio-temporal coherence features, can preserve the spatio-temporal structure of extreme precipitation events.
- 5
- 5
- 5
- 5
- Flow driven attention network for video salient object detection5
Experimental results show that the proposed flow-driven attention network (FDAN) can achieve performance on par with the state-of-the-art method flow-guided recurrent neural encoder on challenging benchmarks of Densely Annotated Video Segmentation and Freiburg-Berkeley Motion Se segmentation while being two times faster in detection.
- Hierarchical Convolutional Neural Network for Face Detection5
The proposed approach of hierarchical convolutional neural network for face detection reaches the state-of-the-art performance and is evaluated on the AFW dataset, FDDB dataset and Pascal Faces, and it has effective generalization.
- 5
- 5
- Registration of Lung Tissue Between Fluoroscope and CT Images: Determination of Beam Gating Parameters in Radiotherapy5
A new registration method on deformable soft tissue between fluoroscopic images and DRR (Digitally Reconstructed Radiograph) images from planning CT images using active shape models is proposed and preliminary results show that the method is very robust for the registration of deformablesoft tissue.
- 5
- 5
- Enhanced Spatiotemporal Consistency for Image-to-LiDAR Data Pretraining4
This work proposes SuperFlow++, a novel framework that integrates spatiotemporal cues in both pretraining and downstream tasks using consecutive LiDAR-camera pairs and establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving.
- 4
- 4
- ReX: An Efficient Approach to Reducing Memory Cost in Image Classification4
A novel approach named recurrent aggregation operator (ReX), which uses recurrent neural networks (RNNs) to effectively aggregate intra-patch features within a large receptive field to get delicate local representations, while bypassing large early activations.
- 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.
- Spatially Regularized Streaming Sensor Selection4
This paper proposes to perform sensor selection in a multi-variate interpolation framework, such that the data sampled by the selected sensors can well predict those of the inactive sensors, and incorporates sensors' spatial information as two regularizers, which leads to significantly better prediction performance.
- Sparse random projection for χ2 kernel linearization: Algorithm and applications to image classification4
A sparse random projection method to linearly approximate the χ2 kernel, so that the original nonlinear SVMs could be converted to linear ones and the improvement comes at almost no cost of accuracy.
- Learning the object location, scale and view for image categorization with adapted classifier4
The experimental results show that the proposed multiple instance learning scheme and the proposed adaptive learning method outperforms the state-of-the-arts in classification accuracy.
- Teacher Agent: A Knowledge Distillation-Free Framework for Rehearsal-Based Video Incremental Learning3
A knowledge distillation-free framework for rehearsal-based video incremental learning called Teacher Agent is proposed, which introduces an agent generator that is either parameter-free or uses only a few parameters to obtain accurate and reliable soft labels.
- 3
- 3
- Valence-engineering modulation of MoS 2 clusters for enhancing biocatalytic activity3
This work develops ultra-small and water-soluble MoS2 clusters with superior antioxidant activity and enzyme-like activity via valence-engineering modulation with Ce doping to improve the biocatalytic activity of MoS2 clusters via valence engineering modulation, which exhibits great potential in the field of biomedical application, especially inflammatory diseases.
- 3
- 3
- PointNet-Based Channel Attention VLAD Network3
A VLAD enhanced Feature Aggregate Module is proposed to aggregate local features adaptively and a Channel Attention Module is applied to the features to reassemble the elements in high-dimension feature space to extract more informative features.
- 3
- 3
- Unifying Subspace and Distance Metric Learning with Bhattacharyya Coefficient for Image Classification3
Inspired by the Bayesian classification error estimation, the objective function is formulated by minimizing the Bhattachyrra coefficient between two distributions and extended with the kernel trick to learn nonlinear distance metric.
- Comprehensive model of mid-and long-term load forecasting based on maximum entropy principle3
A comprehensive forecasting model based on maximum entropy principle in information theory is introduced and indicates that the new comprehensive model improves the forecasting precision and makes the forecasted results more reasonable and impersonal.
- 3
- 2
- 2
- 2
- Vector-angle geometric mapping-based directional importance sampling method for reliability analysis2
- 2
- 2
- 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
- Fast subspace segmentation via Random Sample Probing2
An iterative method called Random Sample Probing (RANSP), which has a computational complexity of O(n) and can therefore handle large-scale datasets, is proposed and experiments confirm the effectiveness and efficiency of RANSP.
- 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.
- 2
- An Immunology-Inspired Fault Detection and Identification System2
This research makes use of immunological concepts to develop a robust fault detection and identification mechanism, capable of detecting and classifying diverse system faults dynamically and adaptive to new types of faults and overlapping faults.
- Ordinal preserving projection2
A novel dimensionality reduction algorithm named ordinal preserving projection (OPP) for learning to rank, which aims at leveraging the global structure of the data set and ordinal information of the observations to obtain two optimal projection matrices.
- Head tracking using shapes and adaptive color histograms2
The method is demonstrated to be a real-time tracker and robust to clutter, scale variation, occlusion, rotation and camera motion, for several test sequences.
- 1
- 1
- 1
- 1
- TopoDiT-3D: Topology-Aware Diffusion Transformer with Bottleneck Structure for 3D Point Cloud Generation1
The proposed TopoDiT-3D is a Topology-Aware Diffusion Transformer with a bottleneck structure utilizing Perceiver Resampler, which not only offers a mode to integrate topological information extracted through persistent homology into feature learning, but also adaptively filters out redundant local features to improve training efficiency.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Facial Components-based Representation for Caricature Face Recognition1
This work proposes a new representation for recognition that is fused by the representation learned from photos, caricatures, and generated faces, and its proposed method performs better than other convolutional neural network (CNN)-based representations.
- Temporal-angular fusion of MODIS and MISR images via sparse representation1
A fusion method to generate frequent time series MISR images that can give full play to their respective advantages in the inversion of surface or atmospheric parameters is proposed.
- 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.
- 1
- Content quality based image retrieval with multiple instance boost ranking1
This paper proposes a new idea for image retrieval, in which image retrieval is regarded as a ranking issue by evaluating image content quality by developing Mean, Max and TopK ranking models developed with Boost learning.
- 1
- Automatic Moving Object Segmentation with Accurate Boundaries1
This paper presents a layer-model based method to segment moving objects from image sequence with accurate boundaries, using Graph Cuts technique to solve the energy minimization problem and extract motion layers.
- 1
- 1
- –
- Blind Quality Enhancement of Compressed Video via Fine-Grained Degradation-Guided Sequential Inference–
A sequential inference strategy that adaptively adjusts the number of artifact reduction stages according to the estimated compression level is introduced, which significantly improves enhancement performance.
- –
- VLM-driven fine-grained semantic regularization for low-light image enhancement–
A novel Cross-Modal Semantic Regularization (CMSR) method that uses the rich semantic priors of a pre-trained vision-language model (VLM) to regularize LLIE, enabling high-fidelity image restoration with improved detail preservation and perceptual quality.
- Open-Vocabulary Object 6D Pose Estimation via Modulated Textual Semantics–
This paper introduces a framework that modulates static textual embeddings into adaptable semantic guidance for RGB-D open-vocabulary 6D pose estimation, and designs a Semantic-Anchored Cross-view Fusion module that aggregates image features from the reference and query views using modulated semantics as an invariant anchor.
- –
- –
- –
- –
- –
- –
- –
- Author response for "Multi-Granularity Aggregation Network for Remote Sensing Few-Shot Segmentation"–
- –
- –
- –
- Human pose transfer via shape-aware partial flow prediction network–
A shape-aware partial flow prediction network to implement human pose transfer and an image completion module, which exploits the correlation of textures to complete the features of invisible regions using that of visible regions, are proposed.
- –
- 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.
- –
- –
- –
- Top-Down Fusing Multi-level Contextual Features for Salient Object Detection–
A novel network architecture is proposed that progressively fuses the rich multi-level contextual features from top to bottom to learn a more effective feature presentation for robust SOD and is superior against state-of-the-arts in terms of all metrics.
- –
- Local Context Embedding Neural Network for Scene Semantic Segmentation–
This paper presents a novel and effective architecture for scene semantic segmentation, named Local Context Embedding (LCE) network, which passes spatially varying contextual information both horizontally and vertically across each small patch derived from fully convolutional feature maps, through the use of Long Short-Term Memory (LSTM) network.
- –
- –
- –
- –
- –
- –
- –
- A Structure-Adaptive Matching Pursuit Subspace Search Algorithm for Effective Image Sparse Representation–
A structure adaptive matching pursuit subspace search algorithm is proposed to obtain effective image sparse representation and the experimental results confirm the efficiency of the algorithm.
- U-Boot based S3C2440 logo design and implementation–
Experiment results show that achieving the boot logo on U-Boot can provide developer and user with friendly boot information.
- –
- –
- –
- –
- A One-Layer Recurrent Neural Network with a Discontinuous Activation Function for Linear Programming–
- –
- Learning LocalDescriptors forFaceDetection–
Experimental results show that the proposed facedetection approach based on local structure and texture of the objects ingray-level images can be used for realtime applications with satisfied detection performance and the computational cost is very low.
- Improving ICA Performance for Modeling Image Appearance with the Kernel Trick–
This paper proposes to combine the nonlinear kernel trick to improve the performance of ICA for modeling image appearance, where the kernel trick is used to project the input data into a high-dimensional implicit feature space, and then ICA is performed in this implicitfeature space to extract nonlinear independent components of input data.
- Research progreses of physical fatigue–
To resolve questions such as resisting fatigue, eliminating fatigue, keep fit, future studies should be on not only assess fatigue in micro, but also on the possible mechanisms by bioelectricity researching.
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 2026-10-14. 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.