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
- Jun Hai GuoSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works274 from public data
- Deep visual domain adaptation: A survey2,407
A taxonomy of different deep domain adaption scenarios according to the properties of data that define how two domains are diverged is presented, and some potential deficiencies of current methods and several future directions are highlighted.
- Deep Facial Expression Recognition: A Survey1,796
This survey provides a comprehensive review of deep FER, including datasets and algorithms that provide insights into these intrinsic problems, and introduces existing novel deep neural networks and related training strategies that are designed for FER based on both static images and dynamic image sequences.
- Deep face recognition: A survey1,485
A comprehensive review of the recent developments on deep FR, covering broad topics on algorithm designs, databases, protocols, and application scenes, and the technical challenges and several promising directions are highlighted.
- 1,411
- 1,056
- 660
- 576
- Mixed High-Order Attention Network for Person Re-Identification415
This paper proposes the High-Order Attention (HOA) module to model and utilize the complex and high-order statistics information in attention mechanism, so as to capture the subtle differences among pedestrians and to produce the discriminative attention proposals.
- 408
- Comments on "Globally Maximizing, Locally Minimizing: Unsupervised Discriminant Projection with Application to Face and Palm Biometrics"387
This communication shows that the basic ideas of UDP and LPP are identical, and UDP is just a simplified version of LPP on the assumption that the local density is uniform.
- Representative Forgery Mining for Fake Face Detection294
This work proposes an attention-based data augmentation framework to guide detector refine and enlarge its attention, and tracks and occludes the Top-N sensitive facial regions, encouraging the detector to mine deeper into the regions ignored before for more representative forgery.
- 293
- 285
- 242
- 222
- 218
- Domain Generalization via Shuffled Style Assembly for Face Anti-Spoofing213
A novel Shuffled Style Assembly Network (SSAN) is proposed to extract and reassemble different content and style features for a stylized feature space and a new large-scale benchmark for FAS is built up to further evaluate the performance of algorithms in reality.
- 197
- 181
- 158
- 148
- Noisy Softmax: Improving the Generalization Ability of DCNN via Postponing the Early Softmax Saturation146
This paper empirically verifies the superiority of the early softmax desaturation, and proposes Noisy Softmax to mitigating this early saturation issue by injecting annealed noise in softmax during each iteration and improves the generalization ability of CNN model by regularization.
- A Deeper Look at Facial Expression Dataset Bias144
A novel deep Emotion-Conditional Adaption Network (ECAN) is proposed to learn domain-invariant and discriminative feature representations, which can match not only the marginal distribution but also the class-conditional distribution across domains by exploring the underlying label information of the target dataset.
- Blended Emotion in-the-Wild: Multi-label Facial Expression Recognition Using Crowdsourced Annotations and Deep Locality Feature Learning139
A new deep manifold learning network is proposed, called Deep Bi-Manifold CNN, to learn the discriminative feature for multi-label expressions by jointly preserving the local affinity of deep features and the manifold structures of emotion labels.
- SwinFace: A Multi-Task Transformer for Face Recognition, Expression Recognition, Age Estimation and Attribute Estimation134
A multi-purpose algorithm for simultaneous face recognition, facial expression recognition, age estimation, and face attribute estimation based on a single Swin Transformer based on a single Swin Transformer is presented.
- SFace: Sigmoid-Constrained Hypersphere Loss for Robust Face Recognition131
Considering the imperfection of training databases, it is proposed that intra-class and inter-class objectives can be optimized in a moderate way to mitigate overfitting problem, and a novel loss function, named sigmoid-constrained hypersphere loss (SFace), is proposed.
- Identity-aware CycleGAN for face photo-sketch synthesis and recognition130
An Identity-Aware CycleGAN (IACycleGAN) model is proposed that applies a new perceptual loss to supervise the image generation network and improves CycleGAN on photo-sketch synthesis by paying more attention to the synthesis of key facial regions, such as eyes and nose, which are important for identity recognition.
- 130
- 112
- 107
- Robust, accurate and efficient face recognition from a single training image: A uniform pursuit approach107
Experimental results show that the robustness, accuracy and efficiency of the proposed UP method compare favorably to the state-of-the-art one sample based methods.
- Meta Balanced Network for Fair Face Recognition104
A novel meta-learning algorithm, called Meta Balanced Network (MBN), which learns adaptive margins in large margin loss such that the model optimized by this loss can perform fairly across people with different skin tones in face recognition.
- 102
- Federated Learning for Face Recognition with Gradient Correction93
A novel idea of correcting gradients from the perspective of backward propagation is explored and a softmax-based regularizer is proposed to correct gradients of class embeddings by precisely injecting a cross-client gradient term.
- Cross-Receptive Focused Inference Network for Lightweight Image Super-Resolution91
A novel Cross-receptive Field Guided Transformer is designed to enable the selection of contextual information required for reconstruction by using a modulated convolutional kernel that understands the current semantic information and exploits the information interaction within different self-attention.
- MMFakeBench: A Mixed-Source Multimodal Misinformation Detection Benchmark for LVLMs83
MMFakeBench is introduced, the first comprehensive benchmark for mixed-source MMD, and MMD-Agent is proposed, a novel approach to integrate the reasoning, action, and tool-use capabilities of LVLM agents, significantly enhancing accuracy and generalization.
- Signal-To-Noise Ratio: A Robust Distance Metric for Deep Metric Learning83
This paper proposes a robust SNR distance metric based on Signal-to-Noise Ratio (SNR) for measuring the similarity of image pairs for deep metric learning and proposes Deep SNR-based Metric Learning (DSML) to generate discriminative feature embeddings.
- 83
- 82
- Fine-grained face verification: FGLFW database, baselines, and human-DCMN partnership82
A Deep Convolutional Maxout Network (DCMN) is developed which aim to tolerate the multi-modal intra-personal variations and distinguish fine-grained localized inter-personal facial details and the experimental results suggest that the proposed DCMN method significantly outperforms current techniques such as Deepface, DeepID2, and VGG-Face.
- FKA-Owl: Advancing Multimodal Fake News Detection through Knowledge-Augmented LVLMs81
This paper proposes FKA-Owl, a novel framework that leverages forgery-specific knowledge to augment LVLMs, enabling them to reason about manipulations effectively and achieves superior cross-domain performance compared to previous methods.
- 76
- 75
- Hybrid-Attention Based Decoupled Metric Learning for Zero-Shot Image Retrieval70
This paper first emphasizes the importance of learning visual discriminative metric and preventing the partial/selective learning behavior of learner in ZSIR, and then proposes the Decoupled Metric Learning (DeML) framework to achieve these individually.
- 69
- Deep face recognition with clustering based domain adaptation68
A new clustering-based domain adaptation method designed for face recognition task in which the source and target domain do not share any classes, which effectively learns discriminative target representation.
- 68
- 67
- 66
- Equidistant prototypes embedding for single sample based face recognition with generic learning and incremental learning64
Although it is fairly simple and parameter-free, LRA, combined with commonly used local descriptors, such as Gabor representation and local binary patterns, outperforms the state-of-the-art methods for several standard experiments on the Extended Yale B, CMU PIE, AR, and FERET databases.
- 63
- Face Feature Extraction: A Complete Review59
This paper believes that a general framework for face feature extraction consists of four major components: filtering, encoding, spatial pooling, and holistic representation, which could be applied in a task with multiple levels.
- OPOM: Customized Invisible Cloak Towards Face Privacy Protection57
This paper proposes a new method, named one person one mask (OPOM), to generate person-specific (class-wise) universal masks by optimizing each training sample in the direction away from the feature subspace of the source identity.
- Disentangling Identity and Pose for Facial Expression Recognition57
This work proposes an identity and pose disentangled facial expression recognition (IPD-FER) model to learn more discriminative feature representation and achieves state-of-the-art recognition performance.
- Face Recognition Using a Unified 3D Morphable Model56
The proposed solution involves a novel approach to learn a subspace spanned by perturbations caused by the missing modes of variation and image degradations, using 3D face data reconstructed from 2D images rather than 3D capture.
- Adversarial Learning With Margin-Based Triplet Embedding Regularization54
This work proposes to improve the local smoothness of the representation space, by integrating a margin-based triplet embedding regularization term into the classification objective, so that the obtained models learn to resist adversarial examples.
- 52
- Large-Scale 3D Shape Retrieval from ShapeNet Core5552
This track aims to provide a benchmark to evaluate large-scale shape retrieval based on the ShapeNet dataset, using ShapeNet Core55, which provides more than 50 thousands models over 55 common categories in total for training and evaluating several algorithms.
- An adaptive AI-based virtual reality sports system for adolescents with excess body weight: a randomized controlled trial51
It is demonstrated that the virtual reality sports therapy could provide an empathetic approach to addressing adolescent obesity, offering comprehensive improvements in physical, psychological and cognitive health beyond mere weight loss.
- Enhancing Generalization of Universal Adversarial Perturbation through Gradient Aggregation51
A simple and effective method called Stochastic Gradient Aggregation (SGA), which alleviates the gradient vanishing and escapes from poor local optima at the same time, and significantly enhances the generalization ability of UAP and outperforms other state-of-the-art methods.
- Efficient Image Super-Resolution With Feature Interaction Weighted Hybrid Network46
This work proposes a Feature Interaction Weighted Hybrid Network (FIWHN), which comprises a series of Wide-residual Distillation Interaction Block (WDIB) as the backbone, and incorporates a Transformer and explores a novel architecture to combine CNN and Transformer.
- The small sample size problem of ICA: A comparative study and analysis46
A locality pursuit approach is introduced which applies the locality preserving projections in the high-dimensional whitened space and performs better than ICA and other conventional approaches, such as Eigenfaces, Laplacianfaces, and Fisherfaces.
- 45
- 45
- 44
- 44
- Point Adversarial Self-Mining: A Simple Method for Facial Expression Recognition41
A simple yet effective approach, called point adversarial self mining (PASM), to improve the recognition accuracy in facial expression recognition (FER) by simulating human learning processes: providing updated learning materials and guidance from more capable teachers.
- 40
- Gabor-Eigen-Whiten-Cosine: A Robust Scheme for Face Recognition40
The novelty of this paper lies in the finding that the same face with complex variations, projected into the Gabor based whitened PCA feature space, is approximately angle invariance.
- APA: Adaptive Pose Alignment for Pose-Invariant Face Recognition39
A new face alignment method for pose-invariant face recognition, called adaptive pose alignment (APA), which can greatly reduce the intra-class difference and correct the noise caused by the traditional method in the alignment process, especially in unconstrained settings is proposed.
- Boosting Facial Expression Recognition by A Semi-Supervised Progressive Teacher38
This paper proposes a semi-supervised learning algorithm named Progressive Teacher (PT) to utilize reliable FER datasets as well as large-scale unlabeled expression images for effective training, which achieves state-of-the-art performance on widely-used databases RAF-DB and FERPlus.
- 37
- A dataset of oracle characters for benchmarking machine learning algorithms36
The Oracle-MNIST dataset, comprising of 28 × 28 grayscale images of 30,222 ancient characters from 10 categories, designed for benchmarking pattern classification, with particular challenges related to image noise and distortion.
- RAF-AU Database: In-the-Wild Facial Expressions with Subjective Emotion Judgement and Objective AU Annotations36
A sign-based and judgement-based approach to annotating blended facial expressions in the wild using crowdsourcing as a promising strategy for labeling in-the-wild facial expressions is developed and a baseline for AU recognition is provided using popular features and multi-label learning methods.
- 33
- 32
- 31
- Targeted Shilling Attacks on GNN-based Recommender Systems30
By automatically constructing the features and edges of the fake users, the proposed framework AutoAttack achieves accurate attacks on a specific group of users while minimizing the impact on non-target users.
- Energy Confused Adversarial Metric Learning for Zero-Shot Image Retrieval and Clustering30
This paper empirically and experimentally demonstrates the importance of learning embedding with good generalization, achieving state-of-the-art performances on the popular CUB, CARS, Stanford Online Products and In-Shop datasets for ZSRC tasks.
- MLFW: A Database for Face Recognition on Masked Faces29
A simple but effective tool to generate masked faces from unmasked faces automatically, and a new database called Masked LFW (MLFW) based on Cross-Age L FW (CALFW) database, which has good visual consistency with the original face.
- 28
- Racial Faces in-the-Wild: Reducing Racial Bias by Deep Unsupervised Domain Adaptation.28
A deep information maximization adaptation network (IMAN) is proposed to bridge the domain gap among different races and the existence of racial bias in FR algorithms, and comprehensive experiments show that the racial bias could be narrowed-down by the algorithm.
- Facial landmark localization by enhanced convolutional neural network28
A new approach to localizing landmarks in facial image by deep convolutional neural network (DCNN) that takes the expectations of the response maps of enhanced CNN and applies auto-encoder model to the global shape vector, which is effective to rectify the outlier points by the prior global landmark configurations.
- Oracle character recognition using unsupervised discriminative consistency network27
A novel unsupervised domain adaptation method for OrCR, which enables to transfer knowledge from labeled handprinted oracle characters to unlabeled scanned data by leveraging pseudo-labeling to incorporate the semantic information into adaptation and constrain augmentation consistency to make the predictions of scanned samples consistent under different perturbations, leading to the model robustness to abrasion, stain and distortion.
- PropagationNet: Propagate Points to Curve to Learn Structure Information26
This paper presents a novel structure-infused face alignment algorithm based on heatmap regression via propagating landmark heatmaps to boundary heatmaps, which provide structure information for further attention map generation and adopts methods from other fields that address the shift variance problem of CNN for face alignment.
- Lighting-aware face frontalization for unconstrained face recognition26
A lighting-aware face frontalization method that aims to generate both lighting-recovered and lighting-normalized frontalized images, based on only five fiducial landmarks, which outperforms the deep learning based methods on the MPIE database under variable pose and lighting.
- Emulating biological strategies for uncontrolled face recognition26
It is shown that a simple system designed by emulating biological strategies of human visual system can largely surpass the state-of-the-art performance on uncontrolled face recognition.
- Depth Map Denoising Network and Lightweight Fusion Network for Enhanced 3D Face Recognition25
An innovative Depth map denoising network (DMDNet) based on the Denoising Implicit Image Function (DIIF) to reduce noise and enhance the quality of facial depth images for low-quality 3D FR.
- 25
- 25
- A high-performance training-free approach for hand gesture recognition with accelerometer25
A high-performance training-free approach for hand gesture recognition with accelerometer that shows better performance than the traditional training-required algorithms of Hidden Markov Model (HMM) and Dynamic Time Warping (DTW).
- Semi-Supervised 2D Human Pose Estimation Driven by Position Inconsistency Pseudo Label Correction Module24
This paper proposes a semi-supervised 2D human pose estimation framework driven by a position inconsistency pseudo label correction module (SSPCM), and introduces an additional auxiliary teacher and uses the pseudo labels generated by the two teacher model in different periods to calculate the inconsistency score and remove outliers.
- 24
- Compressing Fisher Vector for Robust Face Recognition24
This paper puts forward a new Compact FV (CFV) descriptor, and presents a Generalized ITQ (GITQ) method to binarize the authors' CFV, and applies the CFV and GITQ to encode convolutional activations of Convolutional neural networks.
- 23
- From one to many: Pose-Aware Metric Learning for single-sample face recognition23
The Pose-Aware Metric Learning method achieves an 100% accuracy on the Multi-PIE database under the test setting across poses, which is significantly better than the traditional methods that use a large generic image ensemble to learn the cross-pose transformations.
- 22
- 21
- 21
- A Multi-Modal Approach for Driver Gaze Prediction to Remove Identity Bias21
This work proposes an effective multi-model fusion method for driver gaze estimation that achieves state-of-the-art accuracy and ranks 1st at the EmotiW2020 driver gaze prediction sub-challenge.
- 21
- Deep embedding learning with adaptive large margin N-pair loss for image retrieval and clustering19
Adaptive Large Margin N-Pair loss (ALMN) is proposed, where a novel geometrical Virtual Point Generating (VPG) method is proposed to convert a fixed margin into a local-adaptive angular margin, by automatically generating a boundary training sample in feature space.
- 19
- Learning a locality discriminating projection for classification19
A new algorithm called locality discriminating projection (LDP) for subspace learning is introduced, which provides a new scheme for discriminant analysis by considering both the manifold structure and the prior class information.
- Locality preserving and global discriminant projection with prior information19
A new algorithm called locality preserving and global discriminant projection with prior information (LPGDP) for dimensionality reduction and classification, by considering both the manifold structure and thePrior information, where the prior information includes not only the class label but also the misclassification of marginal samples.
- 18
- Unsupervised adaptive hashing based on feature clustering18
A novel hashing method based on feature clustering to simultaneously generate low-dimensional data with balanced variance and preserve the similarity in Euclidean space is explored and has shown its superiority to state-of-the-art methods on three benchmark datasets.
- 18
- 17
- Multi-view Correlation based Black-box Adversarial Attack for 3D Object Detection16
A simple multi-view correlation based adversarial attack method for the camera-LiDAR fusion 3D object detection models and focus on the black-box attack setting which is more practical in real-world systems is proposed.
- 15
- 15
- Gradient Attention Balance Network: Mitigating Face Recognition Racial Bias via Gradient Attention14
A new de-bias method based on gradient attention, called Gradient Attention Balance Network (GABN), which successfully mitigates racial bias in face recognition and learns more balanced performance for people of different races.
- 14
- Comment on "100% Accuracy in Automatic Face Recognition"14
It is argued that the feasibility of image averaging on identification documents is not fully supported by the presented evidence.
- 13
- 13
- 13
- 13
- 13
- 13
- 12
- Improved community structure discovery algorithm based on combined clique percolation method and K-means algorithm12
An improved community structure discovery algorithm named CPMK-Means algorithm, which can achieve better community mining and higher execution efficiency compared with other algorithms, and is superior to other algorithms in terms of precision, recall, accuracy, F-measure and separation.
- 12
- 12
- Fourier Locally Linear Soft Constrained MACE for facial landmark localization12
It is proved that SVM as a classifier can also be used for localization and soft constrained Minimum Average Correlation Energy filter (soft constrained MACE) is proposed, which is more resistent to overfittings to training set than other variants of correlation filter.
- Extended Supervised Descent Method for Robust Face Alignment12
This paper proposes to modify SDM in three respects: multi-scale HOG features are applied orderly as a coarse-to-fine feature detector, global to local constraints of the facial features are considered orderly in regression cascade and rigid Regularization is applied to obtain more stable prediction results.
- Monitoring visual focus of attention via local discriminant projection12
This paper presents a system to monitor a subject's Visual Focus of Attention (VFOA) based on his/her head poses using a local discriminant projection (LDP) method and indicates that this approach outperforms other methods.
- 11
- Identifying Rhythmic Patterns for Face Forgery Detection and Categorization11
A two-stage network for face forgery detection and categorization consisting of a Spatial-Temporal Filter Module (STFM) for PPG signals filtering, and an Adjacency Interaction Module (AIM) for constraint and interaction of P PG signals is proposed.
- Study on feature selection methods in Chinese text categorization11
This paper presents a study of seven feature selection methods that are commonly used in text categorization: document frequency, information gain, mutual information, X\+2 statistic, expected cross entropy, weight of evidence for text, and odds ratio.
- 10
- 10
- 10
- 10
- 9
- Handwritten Chinese character recognition using Local Discriminant Projection with Prior Information9
- Towards Interactive Deepfake Analysis8
This paper introduces a GPT-assisted data construction process resulting in an instruction-following dataset called DFA-Instruct, and constructs an interactive deepfake analysis system called DFA-GPT, as a strong baseline for the community, with the Low-Rank Adaptation (LoRA) module.
- 8
- 8
- 8
- Adaptive Face Recognition Using Adversarial Information Network8
A novel adversarial information network (AIN) is proposed to address the particularity of face recognition, and a graph convolution network is utilized to predict linkage likelihoods between target data and generate pseudo-labels to assist adversarial MI loss.
- H-AT: Hybrid Attention Transfer for Knowledge Distillation8
A novel channel-wise attention transfer loss with a corresponding channel selection method is proposed that shows a complementary effect with other attention based methods, and the hybrid attention transfer achieves state-of-the-art performance on knowledge distillation tasks.
- 8
- Facial Attractiveness Prediction by Deep Adaptive Label Distribution Learning8
This work proposes a novel deep architecture called Deep Adaptive Label Distribution Learning (DALDL), which uses discrete label distribution of possible ratings rather than single label to supervise the learning process of facial attractiveness prediction, and update the label distribution automatically during training process.
- 8
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery7
This paper shows that utilizing 3D priors effectively compensates for structure knowledge deficiencies in 2D priors while preserving the texture details within the restored images, and introduces FREx (Face Restoration under Extreme conditions), a model that combines structure-accurate 3D priors and texture-rich 2D priors in pretrained generative networks for blind face restoration under extreme conditions.
- Cycle label-consistent networks for unsupervised domain adaptation7
Results validate that the proposed method can alleviate the negative influence of falsely-labeled samples and learn more discriminative features, leading to the absolute improvement over source-only model by 9.4% on Office-31 and 6.3% on Image CLEF-DA.
- Generative Model With Coordinate Metric Learning for Object Recognition Based on 3D Models7
This paper proposes a generative model trained with synthetic images rendered from 3D models which can reduce the burden on collecting real training data and make the background conditions more realistic.
- 7
- Distortion Minimization Hashing7
This paper utilizes rate-distortion theory in the hashing quantization procedure and minimize the distortion to reduce the information loss, and presents a quantization-based hashing method named distortion minimization hashing.
- 7
- Robust Discriminant Analysis of Latent Semantic Feature for Text Categorization7
A Discriminative Semantic Feature (DSF) method for vector space model based text categorization that can not only uncover latent semantic structure but also capture the discriminative feature.
- Seek and Solve Reasoning for Table Question Answering6
A Seek-and-Solve pipeline is proposed that instructs the LLM to first seek relevant information and then answer questions, integrating these two stages at the reasoning level into a coherent Seek-and-Solve Chain of Thought (SS-CoT).
- Skeleton2vec: A Self-supervised Learning Framework with Contextualized Target Representations for Skeleton Sequence6
Skeleton2vec is proposed, a simple and efficient self-supervised 3D action representation learning framework, which utilizes a transformer-based teacher encoder taking unmasked training samples as input to create latent contextualized representations as prediction targets.
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- Feature fusion of palmprint and face via tensor analysis and curvelet transform6
The experimental results demonstrate that the proposed multimodal biometric recognition approach outperforms the unimodal solution and the proposed nearly Gaussian fusion (NGF) strategy has a better performance than other fusion rules.
- Locality discriminating indexing for document classification6
Comparative experiments show that the proposed locality discriminating indexing algorithm for document classification isable to derives compact discriminating document representations for classification.
- Prevalence and clustering of cardiovascular risk factors among resident of coastal areas in Qinzhou, Guangxi, China5
A higher prevalence of CRFs exists among adults in Guangxi and several social-economic factors were associated with the presence ofCRFs, suggesting that effective measures to control the CRFs should be implemented, to reduce the risk of cardiovascular disease in adults.
- Selective Pseudo-Labeling with Reinforcement Learning for Semi-Supervised Domain Adaptation5
A reinforcement learning based selective pseudo-labeling method for semi-supervised domain adaptation that demonstrates superior performance to all the comparison methods and proposes a novel target margin loss for the base model training to improve its discriminability.
- Point Adversarial Self Mining: A Simple Method for Facial Expression Recognition in the Wild5
The Point Adversarial Self Mining approach, a simple yet effective way to progressively mine knowledge from training samples, is proposed to produce training data for CNNs to improve the performance and network generality in Facial Expression Recognition task.
- 5
- Omni-supervised Facial Expression Recognition: A Simple Baseline.5
This work proposes to perform omni-supervised learning by exploiting the labeled samples together with a large number of unlabeled data to enhance the robustness of the learned models for various scenarios and adopts a dataset distillation strategy to distill the target task-related knowledge from the new mined samples and compressed them into a very small set of images.
- Deep Local Descriptors with Domain Adaptation5
This work applies maximum mean discrepancy (MMD) criterion to both the fully connected layer and the convolutional layer, which makes the primary local filters of CNN adaptive to the target dataset in an unsupervised manner.
- Adaptive Quotient Image with 3D Generic Elastic Models for Pose and Illumination Invariant Face Recognition5
This paper combines an Adaptive Quotient Image method with 3D Generic Elastic Models (AQI-GEM) to address large pose and illumination variations for face recognition and demonstrates that it outperforms state-of-the-art face recognition methods, with much simpler parameter tuning, and much less training data.
- 5
- 4
- Unsupervised Evaluation of Out-of-distribution Detection: A Data-centric Perspective4
This paper is the first to introduce the unsupervised evaluation problem in OOD detection, which aims to evaluate Ood detection methods in real-world changing environments without OOD labels, and proposes three methods to compute Gscore as an un supervised indicator of OOD Detection performance.
- Dive into the Resolution Augmentations and Metrics in Low Resolution Face Recognition: A Plain yet Effective New Baseline4
This work took a closer look at the impacts of several resolution augmentations and then analyzed the difficulty of LR samples from the perspective of the model gradient produced by different resolution samples, and proposed a more effective Multi-Resolution Augmentation.
- 4
- Transferring discriminative knowledge via connective momentum clustering on person re-identification4
- 4
- CNN-Based Age Classification via Transfer Learning4
This paper adapted contrastive loss to fine-tune the pre-trained VGG-16 over FG-NET to get a better start point and proposed AvgOut-FC Layer to enhance the performance of the models over Audience to demonstrate the effectiveness of the methods.
- Recognizing Compound Emotional Expression in Real-World Using Metric Learning Method4
Real-world facial expressions are re-categories as compound emotional expressions, which can explain human emotions better and a metric learning method as well as several state-of-the-art facial expression classifying methods including SVM are used to recognize them.
- Constrained Spectral Clustering on Face Annotation System4
A design of face annotation system that can generate more informative constraints and better use constraints with constrained spectral clustering is proposed and tested in a lab situation dataset and a real-live dataset.
- Face Recognition via Compact Fisher Vector4
This paper presents a novel approach to make Fisher vector compact and improves its performance, and utilizes handcrafted low-level descriptors as FV do, but retains only 1st order statistics of FV.
- Blur-Robust Face Recognition via Transformation Learning4
A new method for recognizing faces degraded by blur using transformation learning on the image feature, which learns the transformation which both preserves the manifold structure of the original shape images and enhances the class separability, resulting in a wide applications to various descriptors.
- 3
- 3
- Fake-in-Facext: Towards Fine-Grained Explainable DeepFake Analysis3
A unified multi-task learning architecture, FiFa-MLLM, to simultaneously support abundant multimodal inputs and outputs for fine-grained Explainable DeepFake Analysis, can outperform strong baselines on the AGE task and achieve SOTA performance on existing XDFA datasets.
- Enhancing Generalization Of Invisible Facial Privacy Cloak Via Gradient Accumulation3
Gradient Accumulation (GA) is proposed to aggregate multiple small-batch gradients into a one-step iterative gradient to enhance the gradient stability and reduce the usage of quantization operations.
- Confidence-Aware RGB-D Face Recognition via Virtual Depth Synthesis3
This work proposes a domain-independent pre-training framework that utilizes readily available pre-trained RGB and depth models to separately perform face recognition without needing additional paired data for retraining and proposes an innovative Adaptive Confidence Weighting (ACW).
- 3
- Attention distraction with gradient sharpening for multi-task adversarial attack3
Experimental results on the NYUD-V2 and PASCAL datasets demonstrate that the proposed method can improve the generalization ability of adversarial examples among multiple tasks and achieve better attack performance.
- 3
- Regularization techniques for high-dimensional data analysis3
This paper presents a meta-modelling architecture suitable for smart grids that was developed at the Tsinghua National Laboratory for Information Science and Technology (TNList) in China.
- ZigzagNet: Efficient Deep Learning for Real Object Recognition Based on 3D Models3
A 6.2 MB compact parametric model called ZigzagNet based on SqueezeNet is proposed to improve the performance for recognition by applying moving normalization inside micro architecture and adding channel wise convolutional bypass through macro architecture.
- 3
- Real-World Facial Expression Recognition Using Metric Learning Method3
According to experiments, the six-dimension emotion probability vector derived by PR-NCMML is closer to human perception, which leads to better accuracy than the state-of-the-art methods, such as the SVM based algorithms.
- Locally Rejected Metric Learning Based False Positives Filtering for Face Detection3
A locally rejected metric learning (LRML) based false positives filtering method based on affinity propagation clustering algorithm that is better than Viola-Jones face detectors and the combination of the two methods shows an improvement in face detection.
- 3
- 3
- 3
- 3
- Robust Face Recognition from One Training Sample per Person3
The proposed Gabor-based PCA method using Whiten Cosine Similarity Measure for Face Recognition from One training Sample per Person achieves much higher accuracy than the standard Eigenface technique in the authors' large-scale experiment.
- 2
- 2
- Unsupervised Attention Regularization Based Domain Adaptation for Oracle Character Recognition2
A novel unsupervised domain adaptation (UDA) method, i.e., unsupervised attention regularization net?work (UARN), to transfer recognition knowledge from labeled handprinted oracle characters to unlabeled scanned data to achieve the model robustness to flipping.
- SEQ-former: A context-enhanced and efficient automatic speech recognition framework2
A novel ASR framework called SEQ-former is proposed, emphasizing simplicity, simplicity, ciency, and quickness, and it incorporates a Prediction Decoder Network and a Shared Prediction Decoder Network to enhance contextual capabilities.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- Illumination-Recovered Pose Normalization for Unconstrained Face Recognition2
A continuous identity-preserving face pose normalization method is proposed and natural results in terms of preserving the illumination condition of the query face are obtained, based on only five fiducial landmarks.
- 2
- 2
- Evaluation of Smile Detection Methods with Images in Real-World Scenarios2
A systematic study on smile detection with different SVM classifiers finds that HOG features are the most appropriate features to detect smiling faces, which, combined with RBF kernel SVM, achieve an accuracy of \(93.25\,\%\) on GENKI4K database.
- 2
- 2
- Online Regression of Grandmother-Cell Responses with Visual Experience Learning for Face Recognition2
- 2
- Semi-supervised Learning Based on Label Propagation through Submanifold2
A semi-supervised learning algorithm is proposed based on label propagation through submanifold, whose optimal solution can minimize the cost function and satisfy the supervised constraint.
- Gabor Feature Based Classification Using LDA/QZ Algorithm for Face Recognition2
Experimental results show the LDA/QZ algorithm is always the best or comparable to the best in term of recognition accuracy and the state-of-art dimension reduction algorithms and their combinations of Gabor feature for face recognition are performed.
- Device-aware Optical Adversarial Attack for a Portable Projector-camera System1
By incorporating device-aware adaptations into the digital attack algorithm, such as resolution-aware and color-aware adjustments, this paper addresses the limitations of existing projector-camera-based adversarial light attacks in practical FR setups.
- 1
- RGCN-Based Multi-Label Facial Expression Recognition by Combining Semantic and Geometric Information1
- 1
- LWFD: A Simple Light-Weight Network for Face Detection1
A simple light- Weight network for face detection based on the well performance light-weight network backbone which can run on CPU or ARM, that combines a simple but effective detection framework, a hyperparameter to control the number of channels.
- Dual Dominant Local Binary Descriptor for Face Recognition1
The excellent resistance and strong discriminative power under the condition of intermediate to high level of noises, blur and severe lighting on FERET and CAS-PEAL-R1 databases suggests that the DDP algorithm is a strong candidate for local binary face recognition.
- 1
- Learning Facial Point Response for Alignment by Purely Convolutional Network1
This work aims at developing a model that can deal with complex non-linear variations and be invariant to face bounding box distributions, while preserving high alignment accuracy.
- 1
- 1
- Saliency Region Detection via Graph Model and Statistical Learning1
A novel saliency detection algorithm (named as GMSL) via combining graph model and statistical learning is proposed, which generates an initial saliency map by manifold ranking and optimizes it with absorbing Markov chain, both of which are based on graph model.
- 1
- 1
- 1
- 1
- Transformed Principal Gradient Orientation for Robust and Precise Batch Face Alignment1
Experimental results show TPGO based batch alignment is more precise and robust than the state-of-the-art methods such as RASL and SIFT feature base Congealing, and integrated with a SIFT based pre-alignment procedure.
- 1
- 1
- A Clustering Algorithm Based on Adaptive Subcluster Merging1
The results demonstrate that the proposed ASM based clustering algorithm performs better than K-means, complete-link hierarchial, density-based and maximum variance algorithms.
- –
- –
- –
- –
- –
- –
- Dual Information Speech Language Models for Emotional Conversations–
This work disentangles paralinguistic and linguistic information, enabling SLMs to interpret speech through structured representations and preserves contextual understanding by avoiding the generation of task-specific vectors through controlled randomness.
- –
- Reconstruction and Reenactment Separated Method for Realistic Gaussian Head–
A large-scale one-shot gaussian head generator built upon WebSSL and employed a two-stage training approach that significantly enhances the capabilities of generalization and high-frequency texture reconstruction.
- –
- –
- –
- –
- –
- –
- –
- Face Clustering via Adaptive Aggregation of Clean Neighbors–
The proposed novel method named Adaptive Aggregation of Clean Neighbors (AACN) has two stages of preparation before inputting the graph into GCN, which enables nodes to learn more robust features through the GCN module.
- –
- –
- AdaptiveNet: Toward an Efficient Face Alignment Algorithm–
This paper proposes a time-efficient face alignment algorithm while maintain a sufficient algorithmic accuracy, and adopts MobileNet-V2 as its backbone architecture to deal with easy samples, accompanied by a ResNet branch to handle hard examples.
- –
- –
- Robust Supervised Hashing–
A novel solution which uses a maximum and equal Hamming distance code to construct the label information and keeps a comparable accuracy with the state-of-the-art supervised hashing methods.
- –
- Pose Aided Deep Convolutional Neural Networks for Face Alignment–
Pose Aided Convolutional Neural Networks (PACN) is proposed which uses different networks for faces with different poses which achieves better or comparable results than the state-of-the-art methods.
- Intensity Estimation of the Real-World Facial Expression–
This research explores facial intensity estimation, which cares more about the dynamic changes on a face, and uses RankSVM and RankBoost as frameworks to estimate the ranking scores based on sequences of images.
- –
- A New Representation and Identification Approach of the Hybrid-Driven Kinematic Chain–
A set of coding rules of kinematic pairs in one link and matrix representation of multi-pair links is presented, mainly based on the basic principle than shaft must meet pin in rotating pair of k cinematic chain, resulting in mathematical basis of the establishment of plus or minus characteristics of k Cinematic pair.
- Face Recognition Using Local PCA Filters–
This work proposes an efficient feature extraction architecture based on PCANet that performs far better than many traditional artificial feature extraction methods with the help of standalone filter learning and multiscale local feature combination.
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.