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
- Jingyu YangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works277 from public data
- 542
- TokenPose: Learning Keypoint Tokens for Human Pose Estimation418
A novel approach based on Token representation for human Pose estimation (TokenPose), in detail, each keypoint is explicitly embedded as a token to simultaneously learn constraint relationships and appearance cues from images.
- TransPose: Keypoint Localization via Transformer417
This work proposes a model called Trans-Pose, which introduces Transformer for human pose estimation, and the attention layers built in Transformer enable the model to capture long-range relationships efficiently and also can reveal what dependencies the predicted key-points rely on.
- SimCC: A Simple Coordinate Classification Perspective for Human Pose Estimation248
The proposed SimCC uniformly divides each pixel into several bins, thus achieving sub-pixel localization precision and low quantization error, and can omit additional refinement post-processing and exclude upsampling layers under certain settings, resulting in a more simple and effective pipeline for HPE.
- SpatialBot: Precise Spatial Understanding with Vision Language Models220
SpatialBot, a model designed to enhance spatial understanding by utilizing both RGB and depth images, is proposed and the SpatialQA and SpatialQA $\boldsymbol{E}$ datasets are introduced to train VLMs for depth perception.
- Correlation-Aware Deep Tracking200
This work proposes a novel target-dependent feature network inspired by the self-/cross-attention scheme that deeply embeds cross-image feature correlation in multiple layers of the feature network and is able to suppress non-target features, resulting in instance-varying feature extraction.
- 184
- A multi-manifold discriminant analysis method for image feature extraction184
A Multi-Manifold Discriminant Analysis method based on graph embedded learning and under the Fisher discriminant analysis framework is proposed, leading to promising image recognition performance.
- Kernel sparse representation based classification161
A kernel sparse representation based classification (KSRC) algorithm is proposed, which has more powerful classification ability than SRC and is demonstrated to be effective in face recognition, palmprint recognition and finger-knuckle-print recognition.
- A collaborative representation based projections method for feature extraction160
Experimental results on FERET, AR, Yale face databases and the PolyU finger-knuckle-print database demonstrate that CRP works well in feature extraction and leads to a good recognition performance.
- Vision-Based UAV Self-Positioning in Low-Altitude Urban Environments140
A new dataset, DenseUAV, that is the first publicly available dataset tailored for the UAV self-positioning task and incorporates metric learning into representation learning to enhance the model’s discriminative capacity and to reduce the modality discrepancy.
- 130
- The Ninth Visual Object Tracking VOT2021 Challenge Results127
The Visual Object Tracking challenge VOT2021 is the ninth annual tracker benchmarking activity organized by the VOT initiative; results of 71 trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in recent years.
- 121
- Learning robust and discriminative low-rank representations for face recognition with occlusion114
This paper proposes a novel method by exploiting the low-rankness of both the data representation and each occlusion-induced error image simultaneously, by which the global structure of data together with the error images can be well captured.
- Support vector machine-based optimized decision threshold adjustment strategy for classifying imbalanced data112
The proposed strategy is self-adapting and can find the optimal moving distance of the classification hyperplane according to the real distributions of training samples and is extended to develop an ensemble version (EnSVM-OTHR) that can further improve the classification performance.
- 106
- Learning Tracking Representations via Dual-Branch Fully Transformer Networks76
A Siamese-like Dual-branch network based on solely Transformers for tracking, which outperforms the state-of-the-art methods on the GOT-10k and VOT2020 benchmarks and achieves real-time inference speed (about 40 fps) on one GPU.
- Face recognition using adaptive local ternary patterns method70
An adaptive local feature descriptor for face recognition based on an automatic strategy selecting the threshold for LTP and CS-LBP is introduced and a center-symmetric adaptive local ternary pattern (CS-ALTP) feature description method forFace recognition is proposed.
- Feature extraction based on Laplacian bidirectional maximum margin criterion68
A novel criterion, namely Laplacian bidirectional maximum margin criterion (LBMMC), is developed to address the issue of discriminative information within the local structures of samples and the structural information embedding in the images.
- ADNet: Lane Shape Prediction via Anchor Decomposition66
This paper revisits the limitations of anchor-based lane detection methods and introduces the Large Kernel Attention for Feature Pyramid Network (FPN) and the General Lane IoU (GLIoU) loss, which significantly improves the performance of ADNet in complex scenarios.
- 65
- A multi-task Faster R-CNN method for 3D vehicle detection based on a single image57
A 3D vehicle detection method using a single image is proposed to generate the 3D space coordinate information of the object using monocular vision for autonomous driving and is named multi-task Faster R- CNN (MT-Faster R-CNN).
- A local multiple patterns feature descriptor for face recognition57
A local multiple patterns feature descriptor based on the Weber's law for feature extraction and face recognition and a multi-scale block LMP is presented to generate more discriminative and robust visual feature.
- Hourly solar irradiance forecasting based on encoder–decoder model using series decomposition and dynamic error compensation56
The experimental results indicate that the proposed model shows accurate and robust forecasting performance and is a reliable alternative to hourly irradiance forecasting.
- Robust gait recognition using hybrid descriptors based on Skeleton Gait Energy Image55
A new model-based representation, Skeleton Gait Energy Image (SGEI), has been proposed in this paper and the experimental performances indicate that the proposed methods are more robust to the cloth changes, and contribute to increasing the robustness of gait recognition in the unconstrained environments with view variances and clothes changes.
- iVQA: Inverse Visual Question Answering54
This work poses question generation as a multi-modal dynamic inference process and proposes an iVQA model that can gradually adjust its focus of attention guided by both a partially generated question and the answer.
- AL-ELM: One uncertainty-based active learning algorithm using extreme learning machine50
Experimental results show that AL-ELM is better than or at least comparable to several state-of-the-art uncertainty-based active learning algorithms and could effectively decrease the running time of learning procedure.
- 48
- 46
- Kernel Low-Rank Representation for face recognition44
This work proposes a novel image classification method for robust face recognition, named Low-Rank Representation-based Classification (LRRC), based on seeking the lowest-rank representation of a set of test samples with respect to aSet of training samples, which has the natural discrimination to perform classification.
- Clustering-driven unsupervised deep hashing for image retrieval43
A novel end-to-end deep framework for image retrieval, namely Clustering-driven Unsupervised Deep Hashing (CUDH), to recursively learn discriminative clusters by soft clustering model and produce binary code with high similarity responds is proposed.
- 39
- 38
- 37
- A novel pixel neighborhood differential statistic feature for pedestrian and face detection37
The pixel neighborhood differential pattern is learned with both supervised and unsupervised learning methods, which allow discovering discriminative pixel differential patterns in local area and achieving state-of-the-art results.
- 36
- Finger-Knuckle-Print Recognition Using LGBP35
Inspired by the success of Local Gabor Binary Patterns (LGBP) in face recognition, this work presents a method that uses LGBP to identify finger-knuckle-print images and shows that the proposed method works well.
- Feature extraction using fuzzy inverse FDA35
Experimental results on the ORL, FERET face databases and pulse signal database show that the new method outperforms Fisherface, fuzzy Fisherface and inverse Fisher discriminant analysis.
- Video Based Fall Detection Using Human Poses34
A robust fall detection network to recognize fall events using estimated 3D poses is introduced, which increases respective filed and maintains low computation cost by dilated convolutions.
- Correlation-Embedded Transformer Tracking: A Single-Branch Framework33
This work reformulates the two-branch Siamese tracking as a conceptually simple, fully transformer-based Single-Branch Tracking pipeline, dubbed SBT, and proposes an improved tracker dubbed SuperSBT, which greatly raises the speed of SBT and adopts a hierarchical architecture with a local modeling layer to enhance shallow-level features.
- 33
- Dual-verification network for zero-shot learning33
A novel method, namely, dual-verification network (DVN), which accepts features and attributes in a pairwise manner as input and verifies the result in both the attribute and feature spaces, and adopts the concept of semantic feature representation.
- An improved multihop-based localization algorithm for wireless sensor network using learning approach33
This paper proposes a new approach for multihop localization in wireless sensor network based on nonlinear mapping and learning algorithm, and evaluates the algorithm under various anisotropic network and real environment, and demonstrates the superior performance of the proposed algorithm in terms of location estimation adaptability.
- Inverse Visual Question Answering: A New Benchmark and VQA Diagnosis Tool31
This paper proposes a variational iVQA model that can generate diverse, grammatically correct and content correlated questions that match the given answer, and shows that iVZA is an interesting benchmark for visuo-linguistic understanding, and a more challenging alternative to VQA because an iV QA model needs to understand the image better to be successful.
- 30
- 29
- Infrared small target detection using tensor based least mean square27
The experimental results show that TLMS is an effective infrared small target detection method which can well suppress the background and highlight the infrared small targets.
- 26
- Kernel maximum scatter difference based feature extraction and its application to face recognition26
The proposed method first maps the input data into a potentially much higher dimensional feature space by virtue of nonlinear kernel trick, and in such a way, the problem of feature extraction in the nonlinear space is overcome.
- Discovering and Distinguishing Multiple Visual Senses for Polysemous Words25
This work presents a novel framework that solves the problem of polysemy by allowing sense-specific diversity in search results by discovering a list of possible semantic senses and training a visual classifier for each selected semantic sense.
- Human activity prediction using temporally-weighted generalized time warping25
A temporally-weighted GTW (TGTW) algorithm is developed that is combined with the k-nearest neighbors algorithm for predicting the activity class of an input sequence and evaluated on several publicly available datasets in comparison with state-of-the-art approaches.
- Discriminative low-rank dictionary learning for face recognition25
A novel dictionary learning method to learn a discriminative dictionary which is more suitable for face recognition is presented and an ideal regularization term with label information of training data is introduced to obtain low-rank coefficients.
- A novel label learning algorithm for face recognition25
A novel FME algorithm is proposed for face recognition that devices a score fusion scheme to predict the label of the original unknown sample and obtains mirror images of all original face images and views both mirror images andOriginal face images as available samples.
- A Probabilistic Zero-Shot Learning Method via Latent Nonnegative Prototype Synthesis of Unseen Classes23
A probabilistic framework by taking covariance into account to deal with the problem of nonuniform variances between different categories ofZero-shot learning, a type of structured multioutput learning, which suffers from great performance degradation.
- Two-directional maximum scatter difference discriminant analysis for face recognition22
The proposed method not only avoids the singularity problem frequently occurred in the classical Fisher discriminant analysis due to the small sample size, but also saves much computational time.
- SSPNet: Scale and spatial priors guided generalizable and interpretable pedestrian attribute recognition21
A novel Scale and Spatial Priors Guided Network (SSPNet) is proposed for PAR, which is mainly composed of the Adaptive Feature Scale Selection (AFSS) and Prior Location Extraction (PLE) modules and achieves superior performance on the PCS dataset for attribute localization in terms of IoU.
- GPA-3D: Geometry-aware Prototype Alignment for Unsupervised Domain Adaptive 3D Object Detection from Point Clouds21
This work proposes a novel unsupervised domain adaptive 3D detection framework, namely Geometry-aware Prototype Alignment (GPA-3D), which explicitly leverages the intrinsic geometric relationship from point cloud objects to reduce the feature discrepancy, thus facilitating cross-domain transferring.
- Subspace-based multi-view fusion for instance-level image retrieval21
A subspace-based multi-view fusion strategy where a shared subspace is uncovered from the original high-dimensional features yielding a compact latent representation to achieve the state-of-the-art performance.
- Facial Emotion Recognition via Discrete Wavelet Transform, Principal Component Analysis, and Cat Swarm Optimization21
A novel facial emotion recognition method using discrete wavelet transform to extract wavelet coefficients from facial images and introducing the cat swarm optimization to train the weights and biases of the classifier was proposed.
- Low-resolution degradation face recognition over long distance based on CCA21
This work uses canonical correlation analysis to extract the correlation between high-resolution face images and low-resolution ones and to find the transform pair between them, and shows that face images of the same individual with variable resolutions can be matched accurately.
- Intelligence Science and Big Data Engineering21
An advanced Variable window stereo matching method that takes a certain threshold to determine matching windows and a cost function using the non-parametric and gray value is proposed to improve the accuracy of matching further.
- 20
- PCANet: An energy perspective20
This paper tries to explain why PCANet works well from energy perspective point of view based on a set of experiments and finds that this error rate is correlated with the logarithm of energy of image.
- Extracting Privileged Information from Untagged Corpora for Classifier Learning19
A new instance-level multi-instance learning (MIL) model is proposed to enhance classifier learning by extracting PI from untagged corpora, which can effectively eliminate the dependency on manually labeled data.
- A multi-attribute multi-item negotiation mechanism of supply chain networks between buyers and sellers19
A coalition formation-based negotiation protocol is proposed for the multi-attribute multi-item negotiation between multiple manufacturer agents (MAs) and multiple material supplier agents (MSAs).
- Real-time human detection based on gentle MILBoost with variable granularity HOG-CSLBP19
A variable granularity HOG-CSLBP feature, which combines the human shape information with local texture information, and encodes spatial relationship in different granularity to improve its discriminative ability, and a Gentle MILBoost algorithm which utilizes the Newton update technique to get an optimal weak classifier that is able to discriminate complex distribution and is more stable in numerical computation.
- 19
- Sequential Row–Column 2DPCA for face recognition19
For better representation and feature extraction of face images, a two-stage method, called Sequential Row–Column 2D PCA (RC2DPCA), is proposed in this paper, which uses 2DpcA operated in the row direction and alternative 2DPC a operated in column direction.
- DeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation Through Loopback Synergy18
This work proposes DeRIS, a novel framework that decomposes RIS into two key components: perception and cognition, which demonstrates inherent adaptability to both non- and multi-referents scenarios without requiring specialized architectural modifications, enhancing its general applicability.
- 18
- 18
- 18
- 18
- 18
- 18
- A Nuclear Norm Based Matrix Regression Based Projections Method for Feature Extraction18
A NMR-based projections (NMRP) method for feature extraction and recognition of face recognition using the low rank structural information and NMR can be used to build the graph while NMRP is an effective feature extraction method.
- Feature Extraction Using Laplacian Maximum Margin Criterion18
A novel criterion to address the issue of feature extraction by Laplacian Maximum Margin Criterion, which gets the discriminant vectors by maximizing the difference between between-class laplician matrix and within-class LaplACian matrix.
- 17
- 17
- 17
- 17
- 16
- Improving underwater semantic segmentation with underwater image quality attention and muti-scale aggregation attention16
UnderWater SegFormer (UWSegFormer), a transformer-based framework for semantic segmentation of low-quality underwater images, and the Multi-scale Aggregation Attention (MAA) module, which aggregates sets of semantic features at different scales by extracting discriminative information from high-level features, thus compensating for the semantic loss of detail in underwater objects.
- 16
- Capturing the motion of every joint: 3D human pose and shape estimation with independent tokens16
A novel Transformer-based model with a design of independent tokens that uses a temporal model to focus on capturing the rotational temporal information of each joint, which is empirically conducive to preventing large jitters in local parts.
- 16
- 16
- Discriminative Multi-View Privileged Information Learning for Image Re-Ranking16
A discriminative multi-view re-ranking approach in which both the original global image visual contents and the local auxiliary PI features are simultaneously integrated into a unified training framework for generating the latent subspaces with sufficient discriminating power is proposed.
- 16
- 16
- 16
- A Face Detection Method Based on Cascade Convolutional Neural Network15
A new cascaded convolutional neural network method consisting of two main steps is proposed, which achieves improved performance on the FDDB and PASCAL face datasets.
- PropVG: End-To-End Proposal-Driven Visual Grounding with Multi-Granularity Discrimination14
This work proposes PropVG, an end-to-end proposal-based framework that is the first to seamlessly integrate foreground object proposal generation with referential object comprehension without requiring additional detectors, and introduces a Contrastive-based Refer Scoring module.
- 14
- Exploiting textual and visual features for image categorization14
This work forms image selection and classifier learning as a multi-instance learning problem and proposes to solve the employed problem by the cutting-plane algorithm to suppress the search error induced noisy images.
- A new and fast contour-filling algorithm14
The advantages and disadvantages of the conventional filling algorithms are studied, their original ideas are analyzed, and a new filling algorithm is proposed that presents a very simple idea for finding seeds automatically, and limits the scan within the filling regions.
- 13
- Action recognition using direction-dependent feature pairs and non-negative low rank sparse model13
This paper proposes to use direction-dependent feature pairs (DDFP) to represent actions and a novel non-negative low rank sparse model (NLRM) is developed to encode the features and shows the effectiveness of the proposed framework for human action recognition.
- A supervised dictionary learning and discriminative weighting model for action recognition13
A supervised dictionary learning algorithm for action recognition in still images followed by a discriminative weighting model based on Local Fisher Discrimination which takes into account the local manifold structure and discrimination information of local descriptors is proposed.
- Learning Multiple Probabilistic Decisions from Latent World Model in Autonomous Driving12
LatentDriver, a framework models the environment's next states and the ego vehicle's possible actions as a mixture distribution, from which a deterministic control signal is then derived, and achieves expert-level performance on the recently released closed-loop benchmark Waymax.
- Improving Generalized Visual Grounding With Instance-Aware Joint Learning12
InstanceVG is the first framework to simultaneously tackle both GREC and GRES while incorporating instance-aware capabilities into generalized visual grounding, and is the first framework to simultaneously tackle both GREC and GRES while incorporating instance-aware capabilities into generalized visual grounding.
- 12
- 12
- Robust regression based face recognition with fast outlier removal12
The proposed algorithm provides a novel way to tackle the crucial occlusion problem in face recognition by computed from only the clean data, which gives a more discriminant classification rule.
- 11
- Hypergraph Optimization for Salient Region Detection Based on Foreground and Background Queries11
This paper proposes a foreground- and background-queries-based hypergraph optimization framework for salient region detection that combines foreground queries and background queries to uniformly highlight the salient foreground and suppress the non-salient background.
- SERVE: Soft and Equalized Residual VEctors for image retrieval11
This work proposes a novel image signature termed Soft and Equalized Residual VEctors (SERVE) which is more discriminatively formulated and maintains higher accuracy and outperforms existing state-of-the-art methods and provides a promising paradigm for the image retrieval task.
- Learning discriminative shape statistics distribution features for pedestrian detection11
The proposed feature makes use of data distribution which is rich of discriminative information from positive and negative samples, and also utilizes shape statistics which comes from average human template to separate pedestrians from background patches in specific area with shape statistics.
- K-Means Clustering Based on Density for Scene Image Classification11
A density-based k-means clustering that can achieve higher classification accuracy when applied in multiclass scene image classification and has good repeatability.
- A Novel Multihop Range-Free Localization Based on Kernel Learning Approach for the Internet of Things11
This paper proposes a novel IoT localization method: location estimation-kernel partial least squares (LE-KPLS), which has a high localization precision, great stability and strong generalization performance, without having a high requirement of the number of beacons, and it can well adapt to numerous complicated environments.
- 11
- 11
- 10
- 10
- 10
- 10
- 10
- SPA: Spatially Pooled Attributes for image retrieval10
This paper attempts to encode weak spatial information into attribute embedding for effective image retrieval by partitioning the image into regular grids and extracting Classemes attribute vector from each patch, which results in a large pool of Classeme descriptors followed by VLAD aggregation for generating holistic representation.
- Gender classification using 3D statistical models10
The proposed gender classification method consists of three steps: Align the 3D models to get 3D aligned shapes, perform 3D PCA/ICA transformation on the aligned 3D shapes, and do gender classification on the3D Eigenmodels/ICmodels features using SVM.
- Boosting-Like Deep Convolutional Network for Pedestrian Detection10
This paper proposes the boosting-like algorithm to enhance the system convergence stability through adjusting the updating rate according to the classification condition of samples in the training process, and theoretically gives the derivation of the algorithm.
- A novel distribution-based feature for rapid object detection10
Experimental results demonstrate that the proposed features are more discriminative than traditional Haarlike features and multi-block LBP (MBLBP) features and are also robust in monotonous variations of illumination.
- Learning Gabor Features for Facial Age Estimation10
As a single facial representation, Gabor features surprisedly outperform LBP features or even AAM features, and with global/local feature fusion scheme, fused Gabor and AAM or fused LBP and A AM features can achieve significant improvement in age estimation over single feature representation alone.
- EasyChauffeur: A Baseline Advancing Simplicity and Efficiency on Waymax9
This paper introduces EasyChauffeur, a reproducible and effective planner for both imitation learning (IL) and reinforcement learning (RL) on Waymax, a GPU-accelerated simulator, and proposes Ego-Shifting, a new evaluation setting for assessing planners' robustness.
- [CLS] Token is All You Need for Zero-Shot Semantic Segmentation9
This study provides a couple of key discoveries: (i) the global tokens of the text branch in CLIP provide a powerful representation of semantic information and (ii) these text-side [CLS] tokens can be regarded as category priors to guide CLIP visual encoder pay more attention on the corresponding region of interest.
- 9
- 9
- 9
- 9
- Gender Classification via Global-Local Features Fusion9
The experimental results show that the proposed local-global information combination scheme could significantly improve the gender classification accuracy obtained by either local or global features, leading to promising performance.
- Face Detection Using Binary Template Matching and SVM9
This paper presents an efficient approach to achieve fast and accurate face detection in still gray level images by using the structure of eye region as a robust cue to find possible eye pairs.
- 8
- Learning discriminative feature via a generic auxiliary distribution for unsupervised domain adaptation8
A pluggable generic auxiliary distribution (GAD) module for target domain that can iteratively refine the prediction of the target domain samples to increase the separability of the learned features, thereby increasing the distance between features of different categories.
- Hierarchical Representations with Discriminative Meta-filters in Dual Path Network for Tracking8
A novel dual path network with discriminative meta-filters and hierachical representations with superior performance compared with other state-of-the-arts methods is proposed.
- The Structure Transfer Machine Theory and Applications8
This paper proposes a new representation learning method, named Structure Transfer Machine (STM), which enables feature learning process to converge at the representation expectation in a probabilistic way and theoretically shows that such an expected value of the representation (mean) is achievable if the manifold structure can be transferred from the data space to the feature space.
- 8
- Sparsity analysis versus sparse representation classifier8
This paper proposes a new method named Sparse Regression Analysis (SRA) for object representation and recognition, and shows that the discriminative ability of SRC and SRA derives from the fact that the subset which most compactly expresses the input signal is activated in the regression analysis.
- 8
- Collaborative Representation Based Projections for Face Recognition8
This paper develops a collaborative representation based projections (CRP) for face recognition, which is an unsupervised method that works well in feature extraction and leads to good recognition performance.
- 7
- 7
- OPLS-SR: A novel face super-resolution learning method using orthonormalized coherent features7
A novel FSR learning approach based on orthonormalized partial least squares referred to as OPLS-SR, which first learns a latent coherent feature space of low-dimensional HR and LR face embeddings via a recursive optimization, and then super-resolves the LR face images through global face reconstruction and facial detail compensation.
- Multilayer deep features with multiple kernel learning for action recognition7
This paper proposes integrating a novel representation named multilayer deep features (MDF) of both the human region and whole image area into an extended region-aware multiple kernel learning (ER-MKL) framework to learn a robust classifier for fusing human-region MDF and whole- Region MDF.
- 7
- 6
- Research on Retinal Vessel Segmentation Algorithm Based on a Modified U-Shaped Network6
A retinal vessel segmentation algorithm based on a modified U-shaped network is proposed in this paper that can extract multi-scale vascular features and perform segmentation in an end-to-end manner and exhibits better sensitivity and a higher Matthew's correlation coefficient compared to several state-of-the-art methods.
- Attentional Composition Networks for Long-Tailed Human Action Recognition6
A compositional learning based solution for video-based human action recognition that first learns verb-like and preposition-like components, then shuffles these components to generate samples for the tail classes in the feature space to augment the data for the Tail classes.
- 6
- Kernel Collaborative Representation with Regularized Least Square for Face Recognition6
Kernel CRC_RLS is proposed, a kernel collaborative representation based classification with regularized least square by implicitly mapping the sample into high-dimensional space via kernel tricks, leading to promising performance on FERET face database.
- 6
- Ensemble of Global and Local Features for Face Age Estimation6
A feature fusion method to estimate the face age via SVR, which ensembles global feature from Active Appearance Model (AAM) and the local feature from Gabor wavelet transformation.
- 5
- 5
- 5
- 5
- Discriminative Segmentation Tracking Using Dual Memory Banks5
This work presents a novel discriminative segmentation tracking architecture equipped with dual memory banks, i.e., appearance memory bank and spatial memory bank, which outperforms the leading segmentation tracker D3S on two video object segmentation benchmarks DAVIS16 and DAVIS17.
- 5
- Face recognition using fuzzy maximum scatter discriminant analysis5
In this paper, a novel fuzzy linear classifier, called fuzzy maximum scatter difference (FMSD) discriminant criterion, is proposed to extract features from samples, especially deals with outlier samples.
- 5
- 4
- EM-KD: Distilling Efficient Multimodal Large Language Model with Unbalanced Vision Tokens4
EM-KD, a novel paradigm that enhances the Efficient MLLMs with Knowledge Distillation, outperforms prior Efficient MLLMs on both accuracy and efficiency with a large margin, validating its effectiveness.
- MomentSeg: Moment-Centric Sampling for Enhanced Video Pixel Understanding4
A unified framework that jointly optimizes Temporal Sentence Grounding (TSG) and RefVOS, naturally incorporating key moment grounding capability is proposed, and a Moment-Centric Sampling (MCS) strategy that densely samples informative moments while sparsely sampling non-essential frames is designed.
- 4
- Learnable Cross-Scale Sparse Attention Guided Feature Fusion for UAV Object Detection4
A Learnable Cross-scale Sparse Attention (LCSA) guided feature fusion method to improve the performance of UAV object detection and employ a scale-aware loss function to integrate the normalized Wasserstein distance with CIoU in order to improve the incompatibility of IoU for objects with large scale variance.
- 4
- Beyond ITQ: Efficient binary multi-view subspace learning for instance retrieval4
This paper proposes a binary multi-view fusion framework for directly recovering a latent Hamming subspace from the multi- view features for instance retrieval and achieves competitive retrieval performance comparable to the state-of-the-arts and has excellent scalability in large-scale scenario.
- 4
- Hashing with Inductive Supervised Learning4
A new supervised hashing method to generate class-specific hash codes, which uses an inductive process based on the Inductive Manifold Hashing IMH model and leverage supervised information into hash codes generation to address these difficulties and boost the hashing quality.
- 4
- An Improved Method for Oriented Chamfer Matching4
The proposed method was tested on ETHZ shape class dataset and better detection results were acquired and the method called Procrustes Alignment was utilized to combine all the deformable templates into a unified one.
- 4
- 3
- TATrack: Target-oriented adaptive vision transformer for UAV tracking3
TATrack employs an adaptive scoring suspension mechanism, wherein redundant network layers are bypassed when all token scores meet the suspension criteria, thereby enhancing inference speed and strengthening the visual perception of the target, which improves the target orientation and temporal continuity of the predicted position.
- SFFR: Spatial–Frequency Feature Reconstruction for Multispectral Aerial Object Detection3
A novel spatial and frequency feature reconstruction (SFFR) method, which leverages the spatial–frequency feature representation mechanisms of the Kolmogorov–Arnold network to reconstruct complementary representations in both spatial and frequency domains prior to feature fusion.
- 3
- 3
- 3
- 3
- 3
- 3
- Multi-receptive field spatiotemporal network for action recognition3
A multi-receptive field spatiotemporal (MRF-ST) network to effectively model the spatial and temporal information of different receptive fields and can improve the accuracy of action recognition by learning more visual tempos of different actions.
- Exploit CAM by itself: Complementary Learning System for Weakly Supervised Semantic Segmentation3
The Seed Reactivation to help SLM reactivate non-discriminating regions by setting a boundary for activation values, which successively identifies more regions of CAM is proposed, representing a new state-of-the-art among existing WSSS methods.
- Multi-model Network for Fine-Grained Cross-Media Retrieval3
This work proposes a novel multi-modal network for fine-grained cross-media retrieval that is designed as a single feature extraction network for each media to extract unique features for obtaining precise media feature representation.
- 3
- Probabilistic Hypergraph Optimization for Salient Object Detection3
This work proposes a hypergraph based optimization framework for salient object detection to include not only the pairwise but also the higher-order relations among two or more vertices.
- Face Recognition Using A Low Rank Representation Based Projections Method3
Experimental results on FERET, AR, Yale face databases and the PolyU finger-knuckle-print database demonstrate that LRRP works well for face recognition.
- Ensemble Haar and MB-LBP Features for License Plate Detection3
A new license plate detection algorithm is presented, in which, the Haar and MB-LBP features are combined and the updated rules of the sample weights are revised.
- Face Recognition with Multi-scale Block Local Ternary Patterns3
Experiments on ORL and AR databases show that the proposed MB-LTP method significantly outperforms other LBP based face recognition algorithms.
- 3
- 3
- Gender Classification Using the Profile3
This paper proposes a gender classification method using Principal Components Analysis and Independent Components Analysis to extract discriminative features from profile to estimate the face gender via SVM, and presents a robust profile extraction algorithm.
- 3
- 3
- 3
- B-YOLOv8: an improved algorithm for infrared target detection based on UAV aerial images2
The experimental results demonstrate that the B-YOLOv8 can achieve a remarkable balance between efficiency and accuracy, which outperforms the baseline model in terms of mAP50 by 8.2% on the HIT-UAV and 5.1% on the FLIR ADAS V2, while reduce the number of model parameters by 0.3M.
- 2
- 2
- 2
- Illumination Removal via Gaussian Difference L0 Norm Model for Facial Experssion Recognition2
The proposed illumination standardization algorithm uses an iterative L0 Gaussian difference smoothing method, which achieves a more accurate lighting map estimation by reserving the fewest boundaries, and is proven to provide effective lighting processing technical support for the complex face classification problems which require a high degree of preservation of facial texture.
- 2
- 2
- 2
- 2
- Real-Time Human Detection Based on Optimized Integrated Channel Features2
An optimized integrated channel features which can effectively improve the detection performance at the frame rate of 30 fps on images size of 640x480 is proposed.
- Performance Analysis of Matrix Gait Recognition under Linear Interpolation Framework2
A novel gait period detection algorithms based on pseudo-Zernike moments were proposed and Trace transform and Fan-Beam projection were used as instantiation in CASIA(B) gait database to prove the validity of gait recognition framework.
- 2
- 2
- Probabilistic modeling of disparity uncertainty for robust and efficient stereo matching1
A new uncertainty-aware stereo matching framework is proposed that systematically analyze data uncertainty based on the probabilistic distribution of disparity and efficiently estimate model uncertainty without repeated model training.
- 1
- 1
- 1
- 1
- Learning Efficient Transformer Representation for Siamese Tracker to UAV1
This paper proposes an efficient transformer tracker (ET2) with a lightweight network that performs at frame rates far surpassing real-time, achieving the balance of precision and running speed.
- 1
- DMvLNet: deep multiview learning network for blindly accessing image quality1
A unified deep multiview learning network (DMvLNet) is proposed to combine both semantic contents and image distortions for NR-IQA, using ResNet50 as the backbone and design IQA-aware semantic features instead of directly using a recognition-specific CNN model for IQA.
- Separable Batch Normalization for Robust Facial Landmark Localization1
A novel Separable Batch Normalization module with a Cross-protocol Network Training (CNT) strategy for robust facial landmark localization and a novel attention mechanism that assigns different weights to each branch for automatic selection in an effective style are presented.
- 1
- 1
- 1
- 1
- Using LFDA to Learn Subset-Haar-Like Intermediate Feature Weights for Pedestrian Detection1
The experimental results show that the proposed pedestrian detection method using LFDA to learn the weights of Subset-Haar-like intermediate features can achieve lower miss rate and higher precision, compared with other state-of-arts pedestrian detection methods.
- 1
- Local Dual-Cross Ternary Pattern for Feature Representation1
LDCTP is a feature representation inspired by the sole textural structure of human faces that is efficient and only quadruples the cost of computing Local Binary Pattern and Dual-Cross Pattern.
- Supervised Local High-Order Differential Channel Feature Learning for Pedestrian Detection1
A novel supervised local high-order differential channel feature is proposed for fast pedestrian detection and a local structure preserved supervised dimension reduction method which aims to keep the manifold structure of samples in the feature space is proposed.
- Assigning PLS Based Descriptors by SVM in Action Recognition1
This paper applies a multi-class SVM for assignment instead of assigning descriptors to the nearest (Euclidean distance) visual word in traditional Bag of Visual Words (BOVW) framework and uses the K-nearest neighbor algorithm to classify the histogram of visual words.
- A Panoramic Video System Based on Exposure Adjustment and Non-linear Fusion1
This paper proposes a new video stitch method based on the exposure adjustment and nonlinear fusion algorithm, which achieves a much better performance than traditional linear fusion method, especially when there is a big disparity between cameras.
- A Non-negative Low Rank and Sparse Model for Action Recognition1
A novel non-negative low rank and sparse coding model is developed to encode descriptors for action recognition, which is different from previous methods such as spatio-temporal pyramid matching (STPM).
- A Fast Vanishing Point Detection Method in Structured Road1
A fast vanishing point detection method based on texture features is presented for structured road, which uses joint activities of only four Gabor filters to estimate the local dominant orientation at each pixel and votes to estimates the location of vanishing point.
- An Improved Adaptive Weighted LTP Algorithm for Face Recognition Based on Single Training Sample1
An adaptive weighted LTP algorithm with a novel weighted method involving the standard deviation of the sub-images’ feature histogram is proposed for the single training sample per person (SSPP) problem.
- 1
- Computing tasks sectioning and moving scheme in open cloud computing federation based on mobile Agent1
To deal with the huge investment for the peak requirements based cloud platform in a short time, building an open cloud computing federation based on mobile agent is proposed and the computing task sectioning and moving scheme is validated.
- 1
- 1
- –
- Adaptive multimodal visual tracking via parameter-efficient vision transformers–
SwitchTrack is proposed, a unified framework for multimodal tracking with efficient dynamic integration and on-demand modality switching that achieves superior tracking accuracy and efficiency, with significantly lower training and adaptation overhead compared to state-of-the-art approaches.
- –
- LoongTrack: Exploring long-sequence modeling for visual tracking–
This work improves the video scanning approach by proposing causal consistent scanning, which accounts for the causal nature of tracking sequences, and builds a simple tracking pipeline to explore long sequence modeling for capturing the successive information flow under dense contexts.
- –
- VLG-RMOT: Calibration-before-control for end-to-end referring multi-object tracking–
VLG-RMOT is proposed, a calibration-before-control framework that adapts language to the current visual scene before using it for query control, and achieves the best HOTA among the compared end-to-end methods.
- –
- –
- –
- –
- –
- –
- InceptionMamba: An Efficient Hybrid Network with Large Band Convolution and Bottleneck Mamba–
The traditional one-dimensional strip convolutions are replaced by orthogonal band convolutions in the authors' InceptionMamba to achieve cohesive spatial modeling and global contextual modeling can be achieved via a bottleneck Mamba module, facilitating enhanced cross-channel information fusion and enlarged receptive field.
- Shadow-DETR: Alleviating matching conflicts through shadow queries–
The Shadow-DETR substantially boosts existing methods such as DAB-DETR, Deformable-DETR, and DINO while achieving comparable performance with SOTA methods and proposes a loss-balancing strategy to enhance the effectiveness of shadow samples.
- –
- SiamGMF: an improved target tracking model based on UAV aerial images–
A target tracking model based on UAV aerial images called SiamGMF is proposed, which demonstrates superior tracking performance compared with existing methods, indicating that the proposed approach is well-suited for target tracking in the presence of diverse challenging attributes.
- –
- Weakly Supervised Semantic Segmentation via Mamba Image Restoration–
This work treats the CAMs obtained through traditional methods as low-quality images to be restored and design an image restoration method based on the mamba model that effectively improves the quality of CAMs and significantly enhances the final segmentation performance.
- –
- –
- –
- –
- Explicit Attention Network for Face Alignment–
This paper proposes a novel Explicit Attention network for Face Alignment (EAFA), which applies explicit supervision on attention feature maps using facial landmark heatmaps as a simple and direct choice for supervising spatial attention maps.
- –
- –
- A Novel CNN Architecture for Real-Time Point Cloud Recognition in Road Environment–
This work proposes a novel deep CNN architecture for real-time point cloud features extraction by first code the 3D position of point cloud by the index of vertical and horizontal directions, which can be converted into a multi-channel point feature map.
- A Robust Facial Landmark Detector with Mixed Loss–
This work proposes a robust facial landmark detection algorithm based on coordinates regression in an end-to-end training fashion that achieves real-time speed on 300-W with 68 landmarks, which runs at 85 FPS on a Tesla v100 GPU.
- –
- –
- –
- A Novel Discriminative Weighted Pooling Feature for Multi-view Face Detection–
This paper proposes a novel discriminative weighted pooling feature based on the multiple channel maps for multi-view face detection that has some advantages in accuracy or speed when compared with many other state-of-the-art methods.
- Categorization of Plant and Insect Species via Shape Analysis–
A vectorization method to recover clique histogram vectors from the graphs for classification, which permits the use of a codebook approach to represent the graph in terms of a set of codewords that can be used for purposes of support vector machine classification.
- A Pedestrian Detection Method Based on MB_LBP Features and Intersection Kernel SVM–
The experiments show that the proposed approach has higher precision than the classical algorithm HOG+LinearSVM and the HOG_LBP Features Fusion tested on the established benchmarking datasets—INRIA.
- –
- –
- Fusing Discrete Cosine Transform and Multi-level Center-Symmetric Local Binary Pattern Features for Periocular Recognition–
Experimental results on ORL face database, AR face database and Morph periocular database demonstrate that the proposed method outperforms DCT or LBP features forperiocular recognition.
- Boosting Deformable Part Model by Sample Sharing and Outlier Ablation–
The deformable part model is better trained with the proposed method, and the performance of the newly released voc-release 5 [6] model on the challenging VOC 2007 dataset is boosted.
- –
- –
- A weighted location algorithm using statistics median–
A weighted location method based on statistics median is proposed on the basis of analysis of error that can delete outliers and filter random error, improve the precision of the location and coverage rate of nodes.
- –
- –
- –
- –
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.