Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks244 from public data
- Adversarial Domain Adaptation with Domain Mixup520
This paper presents adversarial domain adaptation with domain mixup (DM-ADA), which guarantees domain-invariance in a more continuous latent space and guides the domain discriminator in judging samples' difference relative to source and target domains.
- Deep Industrial Image Anomaly Detection: A Survey514
This paper provides a comprehensive review of deep learning-based image anomaly detection techniques, from the perspectives of neural network architectures, levels of supervision, loss functions, metrics and datasets.
- DSFD: Dual Shot Face Detector506
A novel detection network named Dual Shot face Detector (DSFD) is proposed which inherits the architecture of SSD and introduces a Feature Enhance Module (FEM) for transferring the original feature maps to extend the single shot detector to dual shot detector.
- Rethinking Counting and Localization in Crowds: A Purely Point-Based Framework436
This paper proposes a purely point-based framework for joint crowd counting and individual localization, called Point to Point Network (P2PNet), and proposes a new metric, called density Normalized Average Precision (nAP), to provide more comprehensive and more precise performance evaluation.
- Chained-Tracker: Chaining Paired Attentive Regression Results for End-to-End Joint Multiple-Object Detection and Tracking382
The two major novelties: chained structure and paired attentive regression, make CTracker simple, fast and effective, setting new MOTA records on MOT16 and MOT17 challenge datasets (67.6 and 66.6, respectively), without relying on any extra training data.
- 368
- A Diffusion-Based Framework for Multi-Class Anomaly Detection336
A Difusion-based Anomaly Detection (DiAD) framework for multi-class anomaly detection, which consists of a pixel-space autoencoder, a latent-space Semantic-Guided (SG) network with a connection to the stable diffusion’s denoising network, and a feature-space pre-trained feature extractor.
- Learning Salient Boundary Feature for Anchor-free Temporal Action Localization329
This paper proposes the first purely anchor-free temporal localization method, which is both efficient and effective, and beats all anchor-based and actionness-guided methods with a remarkable margin on THUMOS14, achieving state-of-the-art results, and comparable ones on ActivityNet v1.3.
- Multimodal Industrial Anomaly Detection via Hybrid Fusion297
This paper proposes Multi-3D-Memory (M3DM), a novel multimodal anomaly detection method with hybrid fusion scheme that outperforms the state-of-the-art (SOTA) methods on both detection and segmentation precision on MVTec- 3D AD dataset.
- Rethinking Mobile Block for Efficient Attention-based Models296
This work rethinks lightweight infrastructure from efficient IRB and effective components of Transformer from a unified perspective, extending CNN-based IRB to attention-based models and abstracting a one-residual Meta Mobile Block (MMB) for lightweight model design.
- TEINet: Towards an Efficient Architecture for Video Recognition275
The proposed TEINet can achieve a good recognition accuracy on these datasets but still preserve a high efficiency, and is able to capture temporal structure flexibly and effectively, but also efficient for model inference.
- Surface Representation for Point Clouds250
Umbrella RepSurf surpasses the previous state-of-the-art by a large margin for classification, segmentation and detection on various benchmarks in terms of performance and efficiency, and two variants of Repsurf, Triangular Rep surf and UmbrellaRepSurf inspired by triangle meshes and umbrella curvature in computer graphics are explored.
- Real-IAD: A Real-World Multi-View Dataset for Benchmarking Versatile Industrial Anomaly Detection243
A large-scale, Real-world, and multi-view Industrial Anomaly Detection dataset, named Real- I AD, which contains 150K high-resolution images of 30 different objects, an order of magnitude larger than existing datasets, making it more challenging than previous datasets.
- IFRNet: Intermediate Feature Refine Network for Efficient Frame Interpolation232
This work devise an efficient encoder-decoder based network, termed IFRNet, for fast in-termediate frame synthesizing, which first extracts pyramid features from given inputs, and then refines the bilateral in-Termediate flow fields together with a powerful intermedi-ate feature until generating the desired output.
- Fast Learning of Temporal Action Proposal via Dense Boundary Generator228
An efficient and unified framework to generate temporal action proposals named Dense Boundary Generator (DBG), which draws inspiration from boundary-sensitive methods and implements boundary classification and action completeness regression for densely distributed proposals.
- 225
- Learning by Analogy: Reliable Supervision From Transformations for Unsupervised Optical Flow Estimation218
This work twists the general unsupervised learning pipeline by running another forward pass with transformed data from augmentation, along with using transformed predictions of original data as the self-supervision signal, and introduces a lightweight network with multiple frames by a highly-shared flow decoder.
- To Choose or to Fuse? Scale Selection for Crowd Counting209
This paper proposes a Scale-Adaptive Selection Network (SASNet), which automatically learns the internal correspondence between the scales and the feature levels, and considers the predictions from other feature levels via weighted average, which helps to mitigate the gap between discrete feature levels and continuous scale variation.
- Learning a Few-shot Embedding Model with Contrastive Learning207
This paper proposes a novel contrastive training scheme dubbed infoPatch, exploiting the patch-wise relationship to substantially improve the popular infoNCE and investigates the contrastive learning with Noise Contrastive Estimation in a supervised manner for training a few-shot embedding model.
- 175
- HifiFace: 3D Shape and Semantic Prior Guided High Fidelity Face Swapping173
This work proposes 3D shape-aware identity to control the face shape with the geometric supervision from 3DMM and 3D face reconstruction method and introduces the Semantic Facial Fusion module to optimize the combination of encoder and decoder features and make adaptive blending, which makes the results more photo-realistic.
- Deepfake Generation and Detection: A Benchmark and Survey158
This survey comprehensively reviews the latest developments in deepfake generation and detection, summarizing and analyzing current state-of-the-arts in this rapidly evolving field and benchmarking representative methods on widely adopted datasets to provide a comprehensive and up-to-date evaluation of the most influential published works.
- Efficient multimodal large language models: a survey157
This survey summarizes the timeline of representative efficient MLLMs, the current state of research in structures and strategies, and the applications, and the limitations of current efficient MLLM research and promising future directions are discussed.
- Real3D-AD: A Dataset of Point Cloud Anomaly Detection155
Reg3D-AD, a registration-based 3D anomaly detection method incorporating a novel feature memory bank that preserves local and global representations is proposed, addressing the limitations in the field of high-precision point cloud anomaly detection.
- SoftPatch: Unsupervised Anomaly Detection with Noisy Data144
Comprehensive experiments demonstrate that SoftPatch outperforms the state-of-the-art AD methods on the MVTecAD and BTAD benchmarks and is comparable to those methods under the setting without noise.
- CLIP-AD: A Language-Guided Staged Dual-Path Model for Zero-shot Anomaly Detection144
A framework called CLIP-AD is proposed to leverage the zero-shot capabilities of the large vision-language model CLIP to leverage the zero-shot capabilities of the large vision-language model CLIP, and a Staged Dual-Path model (SDP) is introduced that leverages features from various levels and applies architecture and feature surgery.
- 143
- IM-IAD: Industrial Image Anomaly Detection Benchmark in Manufacturing141
A uniform IM benchmark is proposed, for the first time, to assess how well IAD algorithms perform, which includes various levels of supervision, learning paradigms, memory usage and inference speed, and efficiency.
- Class-Aware Contrastive Semi-Supervised Learning140
This work proposes a general method named Class-aware Contrastive Semi-Supervised Learning (CCSSL), which is a drop-in helper to improve the pseudo-label quality and enhance the model's robustness in the real-world setting and successfully emphasizes clean label learning and simultaneously reduce noisy label learning.
- Learning Dynamic Alignment via Meta-filter for Few-shot Learning134
This paper proposes to learn a dynamic alignment, which can effectively highlight both query regions and channels according to different local support information and facilitates dynamical knowledge adaptation for few-shot learning.
- 128
- MobileMamba: Lightweight Multi-Receptive Visual Mamba Network116
This work proposes the MobileMamba framework, which balances efficiency and performance, and introduces the Multi-Receptive Field Feature Interaction (MRFFI) module, comprising the Long-Range Wavelet Transform-Enhanced Mamba, Efficient Multi-Kernel Depthwise Convolution, and Eliminate Redundant Identity components.
- 115
- 101
- Not All Pixels Are Matched99
A dense contrastive learning framework (DCLNet) is proposed, which performs pixel-to-pixel dense alignment acting on the intermediate representations, rather than the final deep feature, reducing the modality gap from a new perspective.
- DIRL: Domain-Invariant Representation Learning for Generalizable Semantic Segmentation99
The domain-invariant representation is constructed which suppresses the effect of the domain-specific style on the quality and correlation of the features and can enhance the robustness of various backbone networks with little computational cost.
- Transferable Decoding with Visual Entities for Zero-Shot Image Captioning95
ViECap is proposed, a transferable decoding model that leverages entity-aware decoding to generate descriptions in both seen and unseen scenarios and incorporates entity-aware hard prompts to guide LLMs’ attention toward the visual entities present in the image, enabling coherent caption generation across diverse scenes.
- NTIRE 2022 Challenge on Efficient Super-Resolution: Methods and Results94
The NTIRE 2022 challenge was to super-resolve an input image with a magnification factor of ×4 based on pairs of low and corresponding high resolution images and the aim was to design a network for single image super-resolution that achieved improvement of efficiency measured according to several metrics.
- A Survey on Benchmarks of Multimodal Large Language Models90
A comprehensive review of 200 benchmarks and evaluations for MLLMs, focusing on perception and understanding, recognition and reasoning, specific domains, key capabilities, and other modalities, concludes that evaluation should be regarded as a crucial discipline to support the development of MLLMs better.
- 86
- Towards Highly Accurate and Stable Face Alignment for High-Resolution Videos82
A Fractional Heatmap Regression (FHR) for high-resolution video-based face alignment and a novel stabilization loss that contains two terms to address time delay and non-smooth issues, respectively are proposed.
- 81
- Designing One Unified Framework for High-Fidelity Face Reenactment and Swapping74
This paper proposes an effective end-to-end unified framework to achieve both face reenactment and swapping, and sufficiently transfers identity and attribute based on learned disentangled representations to generate high-fidelity faces.
- 73
- Exploring Plain ViT Reconstruction for Multi-class Unsupervised Anomaly Detection73
A novel ViT-based ViTAD structure is instantiated, designed incrementally from both global and local perspectives, and achieves state-of-the-art results and efficiency on MVTec AD, VisA, and Uni-Medical datasets.
- 73
- Ranking Distance Calibration for Cross-Domain Few-Shot Learning70
This work starts from a generic representation pre-trained by a cross-entropy loss and a conventional distance-based classifier, along with an image retrieval view, to employ a re-ranking process to calibrate a target distance matrix by discovering the k-reciprocal neighbours within the task.
- Adversarial Semantic Data Augmentation for Human Pose Estimation69
Semantic Data Augmentation (SDA), a method that augments images by pasting segmented body parts with various semantic granularity, is proposed, which exploits a generative network to dynamiclly predict tailored pasting configuration.
- Applying the fuzzy SERVQUAL method to measure the service quality in certification & inspection industry69
Study shows that the five dimensions of tangible have the biggest gap between the service quality expectations and perceptions and so, the company (CCIC) need to increase investment in tangible aspects in order to improve their service quality.
- Adaptive Assignment for Geometry Aware Local Feature Matching68
AdaMatcher is in-troduce, which first accomplishes the feature correlation and co-visible area estimation through an elaborate feature interaction module, then performs adaptive assignment on patch-level matching while es-timating the scales between images, and finally refines the co- visible matches through scale alignment and sub-pixel regression module.
- 66
- Learning from Noisy Labels with Decoupled Meta Label Purifier65
A novel multi-stage label purifier named DMLP is proposed, which decouples the label correction process into label-free representation learning and a simple meta label purifiers, and can focus on extracting discriminative feature and label correction in two distinctive stages.
- LCTR: On Awakening the Local Continuity of Transformer for Weakly Supervised Object Localization65
A novel framework built upon the transformer, termed LCTR (Local Continuity TRansformer), which targets at enhancing the local perception capability of global features among long-range feature dependencies and a relational patch-attention module (RPAM), which considers cross-patch information on a global basis.
- MixTeacher: Mining Promising Labels with Mixed Scale Teacher for Semi-Supervised Object Detection60
This paper proposes a novel framework that addresses the scale variation problem by introducing a mixed scale teacher to improve pseudo label generation and scale-invariant learning, and proposes mining pseudo labels using score promotion of predictions across scales, which benefits from better predictions from mixed scale features.
- UltraVideo: High-Quality UHD Video Dataset with Comprehensive Captions59
This paper proposes a high-quality open-sourced UHD-4K text-to-video dataset named UltraVideo, and expands Wan to UltraWan-1K/-4K, which can natively generate high-quality 1K/4K videos with more consistent text controllability, demonstrating the effectiveness of the data curation.
- EATFormer: Improving Vision Transformer Inspired by Evolutionary Algorithm59
A novel pyramid EATFormer backbone that only contains the proposed EA-based transformer (EAT) block, which consists of three residual parts to model multi-scale, interactive, and individual information separately and design a task-related head docked with transformer backbone to complete final information fusion more flexibly.
- ColorFormer: Image Colorization via Color Memory Assisted Hybrid-Attention Transformer57
Experimental results show that the proposed automatic image colorization method via color memory assisted hybrid-attention transformer can generate more realistic and semantically matched color images compared with state-of-the-art methods.
- LLaVA-KD: A Framework of Distilling Multimodal Large Language Models54
This work introduces Multimodal Distillation (MDist) to transfer teacher model's robust representations across both visual and linguistic modalities, and Relation Distillation (RDist) to transfer teacher model's ability to capture visual token relationships.
- DiAD: A Diffusion-based Framework for Multi-class Anomaly Detection54
A Difusion-based Anomaly Detection (DiAD) framework for multi-class anomaly detection, which consists of a pixel-space autoencoder, a latent-space Semantic-Guided (SG) network with a connection to the stable diffusion's denoising network, and a feature-space pre-trained feature extractor.
- NeRF-Loc: Visual Localization with Conditional Neural Radiance Field54
A novel visual re-localization method based on direct matching between the implicit 3D descriptors and the 2D image with transformer to improve the localization robustness when domain gap exists between training and testing phases is proposed.
- APANet: Adaptive Prototypes Alignment Network for Few-Shot Semantic Segmentation54
This work presents an adaptive prototype representation by introducing class-specific and class-agnostic prototypes and thus construct complete sample pairs for learning semantic alignment with query features, which effectively enriches feature comparison and helps yield an unbiased segmentation model in the few-shot setting.
- AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection52
This work presents a simple yet effective AdaptCLIP, which supports zero-/few-shot generalization across domains and provides a training-free approach on target domains once trained on a base dataset.
- Harnessing Synthesized Abstraction Images to Improve Facial Attribute Recognition52
This paper builds a dual-path facial attribute recognition network to utilize features from the original face images and facial abstraction images to improve facial attributes recognition, especially the attributes located on small face regions.
- 51
- High-Fidelity Generalized Emotional Talking Face Generation with Multi-Modal Emotion Space Learning51
This paper proposes a more flexible and generalized framework for emotional talking face generation and uses an Aligned Multi-modal Emotion encoder to embed the text, image, and audio emotion modality into a unified space, which inherits rich semantic prior from CLIP.
- Towards Continual Adaptation in Industrial Anomaly Detection50
A novel method named Distribution of Normal Embeddings (DNE), which utilizes the feature distribution of normal training samples from past tasks and not only effectively alleviates catastrophic forgetting in CAD but also can be integrated with SCL methods to further improve their performance.
- Decoupling Classifier for Boosting Few-shot Object Detection and Instance Segmentation49
This paper proposes an embarrassingly simple but effective method that decouples the standard classifier into two heads that are capable of independently addressing clear positive samples and noisy negative samples which are caused by the missing label.
- ASFD: Automatic and Scalable Face Detector49
This work proposes to search an effective FAE architecture, termed AutoFAE by a differentiable architecture search, which outperforms all existing FAE modules in face detection with a considerable margin, and builds a supernet, which automatically obtains a family of detectors under the different complexity constraints.
- Context-Aware Image Inpainting with Learned Semantic Priors48
A context-aware image inpainting model, which adaptively integrates global semantics and local features in a unified image generator that achieves the state of the art on Places2, CelebA, and Paris StreetView datasets is proposed.
- Uniformity in Heterogeneity: Diving Deep into Count Interval Partition for Crowd Counting47
This work proposes a novel count interval partition criterion called Uniform Error Partition (UEP), which always keeps the expected counting error contributions equal for all intervals to minimize the prediction risk, and proposes a simple yet effective model termed UEPNet, which achieves state-of-the-art performance on several challenging datasets.
- 45
- 43
- Learning Distinctive Margin toward Active Domain Adaptation42
This work proposes a concise but effective ADA method called Select-by-Distinctive-Margin (SDM), which consists of a maximum margin loss and a margin sampling algorithm for data selection and provides theoretical analysis to show that SDM works like a Support Vector Machine.
- SiamRCR: Reciprocal Classification and Regression for Visual Object Tracking42
A novel siamese tracking algorithm called SiamRCR is proposed, addressing this problem with a simple, light and effective solution, and builds reciprocal links between classification and regression branches, which can dynamically re-weight their losses for each positive sample.
- 40
- Self-supervised Feature Adaptation for 3D Industrial Anomaly Detection40
Early multi-modal approaches that attempted to utilize models pre-trained on large-scale visual datasets, i.e., ImageNet, to construct feature databases are investigated, and it is found that directly using these pre-trained models is not optimal, it can either fail to detect subtle defects or mistake abnormal features as normal ones.
- 40
- M3DM-NR: RGB-3D Noisy-Resistant Industrial Anomaly Detection via Multimodal Denoising38
A novel noise-resistant M3DM-NR framework to leveraging strong multi-modal discriminative capabilities of CLIP, which outperforms state-of-the-art methods in 3D-RGB multi-modal noisy anomaly detection.
- CAR: Controllable Autoregressive Modeling for Visual Generation36
This study introduces Controllable AutoRegressive Modeling (CAR), a novel, plug-and-play framework that integrates conditional control into multi-scale latent variable modeling, enabling efficient control generation within a pre-trained visual autoregressive model.
- NTIRE 2021 Challenge on Video Super-Resolution36
This paper presents evaluation results from two competition tracks as well as the proposed solutions to the NTIRE 2021 Challenge on Video Super-Resolution, and develops conventional video SR methods focusing on the restoration quality.
- Learning to Aggregate and Personalize 3D Face from In-the-Wild Photo Collection36
A novel Learning to Aggregate and Personalize (LAP) framework for unsupervised robust 3D face modeling that recovers superior or competitive face shape and texture compared with state-of-the-art (SOTA) methods with or without prior and supervision.
- Temporal Distinct Representation Learning for Action Recognition36
This paper designs a sequential channel filtering mechanism, i.e., Progressive Enhancement Module (PEM), to excite the discriminative channels of features from different frames step by step to avoid repeated information extraction and achieves visible improvements over the best competitor by 2.4% and 1.3%, respectively.
- Learning with Noisy labels via Self-supervised Adversarial Noisy Masking35
This paper empirically finds that models trained with clean and mislabeled samples manifest distinguishable activation feature distributions and proposes a novel robust training approach termed adversarial noisy masking, which adaptively modulates the input data and label simultaneously, preventing the model to overfit noisy samples.
- VividPose: Advancing Stable Video Diffusion for Realistic Human Image Animation34
This work proposes VividPose, an innovative end-to-end pipeline based on Stable Video Diffusion (SVD) that ensures superior temporal stability and enhances the retention of human identity, and introduces a geometry-aware pose controller that utilizes both dense rendering maps from SMPL-X and sparse skeleton maps.
- 33
- PatchMix Augmentation to Identify Causal Features in Few-Shot Learning33
A novel data augmentation strategy dubbed as PatchMix that can break this spurious dependency between patch-level information and supervision of the query images with random gallery images from different classes from the query ones is proposed and theoretically shows that such an augmentation mechanism, different from existing ones, is able to identify the causal features.
- Stroke-based Neural Painting and Stylization with Dynamically Predicted Painting Region32
Compositional Neural Painter is proposed, a novel stroke-based rendering framework which dynamically predicts the next painting region based on the current canvas, instead of dividing the image plane uniformly into painting regions.
- SoftPatch+: Fully unsupervised anomaly classification and segmentation31
This paper is the first to consider fully unsupervised industrial anomaly detection (i.e., unsupervised AD with noisy data) with memory-based unsupervised AD methods, SoftPatch and SoftPatch+, which efficiently denoise the data at the patch level.
- Learning Feature Inversion for Multi-class Anomaly Detection under General-purpose COCO-AD Benchmark29
A simple but more powerful InvAD framework to achieve high-quality feature reconstruction and improves the effectiveness of reconstruction-based methods on popular MVTec AD, VisA, and the newly proposed COCO-AD datasets under a multi-class unsupervised setting.
- Reference Twice: A Simple and Unified Baseline for Few-Shot Instance Segmentation28
A unified framework, Reference Twice (RefT), is introduced to exploit the relationship between support and query features for FSIS and related tasks, demonstrating that support object queries encode key factors after base training, allowing query features to be enhanced twice at both feature and query levels.
- tSF: Transformer-Based Semantic Filter for Few-Shot Learning28
The proposed tSF redesigns the inputs of a transformer-based structure by a semantic filter, which not only embeds the knowledge from whole base set to novel set but also filters semantic features for target category.
- PVG: Progressive Vision Graph for Vision Recognition27
A Progressive Vision Graph (PVG) architecture for vision recognition task is proposed, which contains three main components: Progressively Separated Graph Construction (PSGC) to introduce second-order similarity, Neighbor nodes information aggregation and update module by using Max pooling and mathematical Expectation to aggregate rich neighbor information.
- Phasic Content Fusing Diffusion Model with Directional Distribution Consistency for Few-Shot Model Adaption26
A novel phasic content fusing few-shot diffusion model with directional distribution consistency loss, which targets different learning objectives at distinct training stages of the diffusion model, and a cross-domain structure guidance strategy that enhances structure consistency during domain adaptation.
- StyleFace: Towards Identity-Disentangled Face Generation on Megapixels26
This work proposes StyleFace, a unified framework for 1024 2 resolution high-fidelity identity swapping and de-identification, and proposes an Adaptive Attribute Extractor, which adaptively preserves the identity-irrelevant attributes in a simple yet effective way.
- MMAD: A Comprehensive Benchmark for Multimodal Large Language Models in Industrial Anomaly Detection25
- 25
- 25
- 24
- 23
- SSCGAN: Facial Attribute Editing via Style Skip Connections23
A style skip connection based generative adversarial network, referred to as SSCGAN, is developed which enables accurate facial attribute manipulation and can produce better results in terms of attribute generation accuracy and image quality.
- Real-IAD D 3 : A Real-World 2D/Pseudo-3D/3D Dataset for Industrial Anomaly Detection22
This work introduces Real-IAD D3, a high-precision multimodal dataset that uniquely incorporates an additional pseudo-3D modality generated through photometric stereo, alongside high-resolution RGB images and micrometer-level 3D point clouds, providing a challenging benchmark for multimodal IAD.
- 22
- Dense Scene Multiple Object Tracking with Box-Plane Matching22
The Box-Plane Matching method is proposed to improve the MOT performacne in dense scenes and achieve data association according to the motion similarity and appearance similarity between tracklets and detections.
- 21
- Better “CMOS” Produces Clearer Images: Learning Space-Variant Blur Estimation for Blind Image Super-Resolution21
A novel Cross-MOdal fuSion network (CMOS) that estimate both blur and semantics simultaneously, which leads to improved SR results and involves a feature Grouping Interactive Attention (GIA) module to make the two modalities in-teract more effectively and avoid inconsistency.
- FRIH: Fine-grained Region-aware Image Harmonization21
This work proposes a novel global-local two stages framework for Fine-grained Region-aware Image Harmonization (FRIH), and designs a fusion prediction module to generate the final result, utilizing the different degrees of harmonization results comprehensively.
- 20
- MambaGesture: Enhancing Co-Speech Gesture Generation with Mamba and Disentangled Multi-Modality Fusion19
This work presents MambaGesture, a novel framework integrating a Mamba-based attention block, MambaAttn, with a multi-modality feature fusion module, SEAD, achieving state-of-the-art performance in co-speech gesture generation.
- STC: Spatio-Temporal Contrastive Learning for Video Instance Segmentation19
This work presents a simple and efficient single-stage VIS framework based on the instance segmentation method CondInst by adding an extra tracking head and proposes a novel bi-directional spatio-temporal contrastive learning strategy for tracking embedding across frames.
- ClotheDreamer: Text-guided garment generation with 3D gaussians18
This work introduces ClotheDreamer, a 3D Gaussian-based method for generating wearable, production-ready 3D garment assets from text prompts, and proposes a novel representation Disentangled Clothe Gaussian Splatting (DCGS) to enable separate optimization.
- 3QNet18
A novel learning-based point cloud compression framework named 3D Point Cloud Geometry Quantiation Compression Network (3QNet) is proposed, which overcomes the robustness limitation of existing point-based methods and can handle dense points.
- Towards One-step Causal Video Generation via Adversarial Self-Distillation17
This work proposes a distillation-based framework for efficient causal video generation that enables high-quality synthesis with extremely limited denoising steps and proposes a novel Adversarial Self-Distillation (ASD) strategy, which aligns the outputs of the student model's n-step denoising process with its (n+1)-step version at the distribution level.
- 17
- RealTalk: Real-time and Realistic Audio-driven Face Generation with 3D Facial Prior-guided Identity Alignment Network16
A novel generalized audio-driven framework RealTalk, which consists of an audio-to-expression transformer and a high-fidelity expression-to-face renderer, and a lightweight facial identity alignment (FIA) module which includes a lip-shape control structure and a face texture reference structure is proposed.
- 16
- Learning Global-aware Kernel for Image Harmonization16
A novel Global-aware Kernel Net-work (GKNet) is proposed to harmonize local regions with comprehensive consideration of long-distance background references for local harmonization and a novel Selective Correlation Fusion (SCF) module is proposed to better select relevant long-distance background references for local harmonization.
- Joint Learning Content and Degradation Aware Feature for Blind Super-Resolution16
The proposed Content and Degradation aware SR Network dubbed CDSR outperforms the existing UDP models and achieves competitive performance on PSNR and SSIM even compared with the state-of-the-art SKP methods.
- Disentangle Identity, Cooperate Emotion: Correlation-Aware Emotional Talking Portrait Generation15
Qualitative results and user studies further confirm the DICE-Talk method's ability to generate identity-preserving portraits with rich, correlated emotional expressions that naturally adapt to unseen identities, outperforming state-of-the-art approaches in emotion accuracy while maintaining competitive lip-sync performance.
- Toward High Quality Facial Representation Learning15
A self-supervised pre-training framework, called Mask Contrastive Face (MCF), with mask image modeling and a contrastive strategy specially adjusted for face domain tasks is proposed, which outperforms the state-of-the-art methods on multiple downstream tasks.
- CtlGAN: Few-shot Artistic Portraits Generation with Contrastive Transfer Learning15
This work proposes CtlGAN, a new few-shot artistic portraits generation model with a novel contrastive transfer learning strategy that significantly outperforms state-of-the-arts under 10-shot and 1-shot settings and generates high quality artistic portraits.
- DMAD: Dual Memory Bank for Real-World Anomaly Detection14
The results show that DMAD surpasses current state-of-the-art methods, highlighting DMAD's capability in handling the complexities of real-world anomaly detection scenarios.
- MDT-A2G: Exploring Masked Diffusion Transformers for Co-Speech Gesture Generation14
This work introduces a novel Masked Diffusion Transformer for co-speech gesture generation, referred to as MDT-A2G, which directly implements the denoising process on gesture sequences and proposes an efficient inference strategy that diminishes the denoising computation by leveraging previously calculated results, thereby achieving a speedup with negligible performance degradation.
- Toward Multi-class Anomaly Detection: Exploring Class-aware Unified Model against Inter-class Interference14
A Multi-class Implicit Neural representation Transformer for unified Anomaly Detection (MINT-AD), which leverages the fine-grained category information in the training stage to mitigate inter-class interference within the reconstruction model.
- 14
- A Generalist FaceX via Learning Unified Facial Representation14
This work presents FaceX framework, a novel facial generalist model capable of handling diverse facial tasks simultaneously, and introduces Facial Omni-Representation Decomposing (FORD) for seamless manipulation of various facial components, microscopically decomposing the core aspects of most facial editing tasks.
- 14
- 13
- HiFiVFS: High Fidelity Video Face Swapping13
A high fidelity video face swapping framework which leverages the strong generative capability and temporal prior of Stable Video Diffusion (SVD) and builds a fine-grained attribute module to extract identity-disentangled and fine-grained attribute features through identity desensitization and adversarial learning.
- Frequency Consistent Adaptation for Real World Super Resolution13
A novel Frequency Consistent Adaptation (FCA) that ensures the frequency domain consistency when applying existing Super-Resolution methods to the real scene and improves the performance of the SR model under real-world setting achieving state-of-the-art results with high fidelity and plausible perception.
- 13
- OracleFusion: Assisting the Decipherment of Oracle Bone Script with Structurally Constrained Semantic Typography12
A novel twostage semantic typography framework that outperforms state-of-the-art baseline models in terms of semantics, visual appeal, and glyph maintenance, and provides expertlike insights on unseen oracle characters, making it a valuable tool for advancing the decipherment of OBS.
- 12
- 12
- 12
- Learning Comprehensive Motion Representation for Action Recognition12
A Comprehensive Motion Representation (CMR) learning method for action recognition, which achieves competitive performance on Something-Something V1 & V2 and Kinetics-400, and outperforms the current state-of-the-art on the temporal reasoning datasets Something- something V1 and V2.
- NoiseBoost: Alleviating Hallucination with Noise Perturbation for Multimodal Large Language Models11
NoiseBoost is proposed, a broadly applicable and simple method for alleviating hallucinations for MLLMs through the integration of noise feature perturbations and pioneerly enables semi-supervised learning for MLLMs, unleashing the power of unlabeled data.
- 11
- A MADM Location Privacy Protection Method Based on Blockchain11
A credit value is introduced, credit data from users is converted into credit values using the multiple-attribute decision making (MADM) algorithm, credit values and transaction information from the anonymous zone construction process are stored in conjunction with a blockchain, and a credit value reward and punishment mechanism is used to constrain undesirable behaviors.
- 11
- 11
- Similarity metric learning for face verification using sigmoid decision function11
The proposed face verification method is based on the idea that it would like the similarity between positive pairs larger than negative pairs, and obtain a similarity estimation of two images, and belongs to nonlinear metric which is more robust to handle heterogeneous data than linear metric.
- Real-IAD Variety: Pushing Industrial Anomaly Detection Dataset to a Modern Era10
This work introduces Real-IAD Variety, the largest and most diverse IAD benchmark, and demonstrates that zero-shot and few-shot IAD models exhibit remarkable robustness to category scale-up, maintaining consistent performance and significantly enhancing generalization across diverse industrial contexts.
- 10
- 10
- LLaVA-VSD: Large Language-and-Vision Assistant for Visual Spatial Description10
A Large Language-and-Vision Assistant for Visual Spatial Description, named LLaVA-VSD, which is designed for the classification, description, and open-ended description of visual spatial relationships and demonstrates excellent multimodal conversational capabilities.
- 10
- Textual Decomposition Then Sub-motion-space Scattering for Open-Vocabulary Motion Generation10
This paper proposes to leverage the atomic motion as an intermediate representation, and leverage two orderly coupled steps, i.e., Textual Decomposition and Sub-motion-space Scattering, to address the full mapping problem of the open-vocabulary motion generation.
- 10
- Global Meets Local: Effective Multi-Label Image Classification via Category-Aware Weak Supervision10
This paper builds a unified framework to perform effective noisy-proposal suppression and to interact between global and local features for robust feature learning, and proposes category-aware weak supervision to concentrate on non-existent categories so as to provide deterministic information for local feature learning.
- 10
- ID-Sculpt: ID-aware 3D Head Generation from Single In-the-wild Portrait Image9
This work incorporates the identity information of the portrait image into three parts: 1) geometry initialization, 2) geometry sculpting, and 3) texture generation stages, which can generate high-quality 3D heads with accurate geometry and texture from a single in-the-wild portrait image.
- Dynamic Frame Interpolation in Wavelet Domain9
This work proposes a novel two-stage frame interpolation framework termed WaveletVFI and finds that embedding a classifier in the motion perception network to learn a dynamic threshold for each sample can achieve more computation reduction with almost no loss of accuracy.
- 9
- Rethinking the Metric in Few-shot Learning9
This paper investigates the contributions of different distance metrics, proposes an adaptive fusion scheme, and designs a few-shot classification framework AMTNet, including the AMM and the Global Adaptive Loss, to jointly optimize the few- shot task and auxiliary self-supervised task, making the embedding features more robust.
- 9
- SVFR: A Unified Framework for Generalized Video Face Restoration8
A unified framework, termed as stable video face restoration (SVFR), is presented, which leverages the generative and motion priors of Stable Video Diffusion (SVD) and incorporates task-specific information through a unified face restoration framework.
- 8
- 8
- 8
- Adversarial Refinement Network for Human Motion Prediction8
This work proposes an Adversarial Refinement Network (ARNet) following a simple yet effective coarse-to-fine mechanism with novel adversarial error augmentation, which outperforms other state-of-the-art methods, especially on challenging aperiodic actions in both short-term and long-term predictions.
- Reasoning to Edit: Hypothetical Instruction-Based Image Editing with Visual Reasoning7
This work proposes Reason50K, a large-scale dataset specifically curated for training and evaluating hypothetical instruction reasoning image editing, along with ReasonBrain, a novel framework designed to reason over and execute implicit hypothetical instructions across diverse scenarios.
- 7
- 7
- 7
- 7
- LSTC: Boosting Atomic Action Detection with Long-Short-Term Context7
This paper place the atomic action detection problem into a Long-Short Term Context (LSTC) to analyze how the temporalreliance among video signals affect the action detection results, and demonstrates that both temporal grains are beneficial to atomic action recognition.
- 7
- PSPU: Enhanced Positive and Unlabeled Learning by Leveraging Pseudo Supervision6
This work introduces a pseudo-supervised PU learning framework (PSPU), in which the PU model is train first, use it to gather confident samples for the pseudo supervision, and then apply these supervision to correct the PU model’s weights by leveraging non-PU objectives.
- DynamicControl: Adaptive Condition Selection for Improved Text-to-Image Generation6
This work proposes a novel framework, DynamicControl, which supports dynamic combinations of diverse control signals, allowing adaptive selection of different numbers and types of conditions, and demonstrates its superiority over existing methods in terms of controllability, generation quality and composability under various conditional controls.
- 6
- 6
- 6
- 6
- Hierarchical Walking Transformer for Object Re-Identification6
A new hierarchical walking attention is presented, which provides a scalable, flexible, and interpretable sparsification strategy to reduce the complexity from quadratic to linear, and meanwhile evidently boost the performance.
- 5
- Identity-Preserving Text-to-Video Generation Guided by Simple yet Effective Spatial-Temporal Decoupled Representations5
A simple yet effective spatial-temporal decoupled framework that decomposes representations into spatial features for layouts and temporal features for motion dynamics, and secures the runner-up position in 2025 ACM Multimedia Challenge.
- Learning Multi-view Anomaly Detection with Efficient Adaptive Selection5
The Multi-View Anomaly Detection (MVAD) framework is introduced, which learns and integrates features from multi-views and a Multi-View Adaptive Selection (MVAS) algorithm for feature learning and fusion across multiple views is proposed.
- 5
- 5
- 5
- 5
- HifiHead: One-Shot High Fidelity Neural Head Synthesis with 3D Control5
HifiHead is proposed, a high fidelity neural talking head synthesis method, which can well preserve the source image's appearance and control the motion flexibly with 3D morphable face models (3DMMs) parameters derived from a driving image or indicated by users.
- 4
- 4
- 4
- Enhancing Grid-Based 3D Object Detection in Autonomous Driving With Improved Dimensionality Reduction4
A new point cloud backbone network for grid-based object detection, MDRNet, which is based on adaptive dimensionality reduction and multi-level spatial residual strategies, and designed to dynamically concentrate on the essential components of the object during 3D-to-BEV transformation.
- 4
- 4
- 4
- 4
- CustAny: Customizing Anything from A Single Example3
This paper proposes a novel pipeline to construct a large dataset of general objects and build the Multi-Category ID-Consistent (MC-IDC) dataset, and introduces Customizing Anything (CustAny), a zero-shot framework that maintains ID fidelity and supports flexible text editing for general objects.
- ArtWeaver: Advanced Dynamic Style Integration via Diffusion Model3
A novel framework dubbed StyleMaster is proposed for this task by leveraging pretrained Stable Diffusion (SD), which addresses previous problems such as insufficient style and inconsistent semantics, and introduces two objective functions to optimize the model together with denoising loss, which can further enhance semantic and style consistency.
- 3
- HCPM: Hierarchical Candidates Pruning for Efficient Detector-Free Matching3
HCPM is an efficient and detector-free local feature-matching method that employs hierarchical pruning to optimize the matching pipeline and significantly surpasses existing methods in terms of speed while maintaining high accuracy.
- 3
- 3
- 3
- Learning Versatile 3D Shape Generation with Improved AR Models3
The Improved Auto-regressive Model (ImAM) is proposed for 3D shape generation, which applies discrete representation learning based on a latent vector instead of volumetric grids and can synthesize diverse and faithful shapes of multiple categories, achieving state-of-the-art performance.
- 3
- Evaluating the Quality of Face Alignment without Ground Truth3
This study aims to develop a method that can select an input image with good face alignment results from many results produced by a single method or multiple ones, which can enhance the overall performance of a face alignment method with a random strategy.
- 3
- 2
- 2
- Enhancing Multi-Class Anomaly Detection via Diffusion Refinement with Dual Conditioning2
This approach leverages diffusion to obtain high-frequency information for refinement and introduces category-awareness using dual conditioning to ensure the accuracy of prediction and reconstruction in the reverse diffusion process, preventing excessive deviation from the target category, thus effectively enabling multi-class anomaly detection.
- Decision Boundary-aware Knowledge Consolidation Generates Better Instance-Incremental Learner2
A novel decision boundary-aware distillation method with consolidating knowledge to teacher to ease the student learning new knowledge and demonstrates that the teacher model can be a better incremental learner than the student model, which overturns previous knowledge distillation-based methods.
- 2
- Dual Path Transformer with Partition Attention2
This paper introduces a novel attention mechanism, called dual attention, which is both efficient and effective, and presents a hierarchical vision backbone called DualFormer, which achieves 81.5\% top-1 accuracy on ImageNet, outperforming the recent state-of-the-art MPViT-XS by 0.6%, and evaluates the effectiveness of the model on several computer vision tasks.
- 2
- Delving into Transformer for Incremental Semantic Segmentation2
This work proposes TISS, a Transformer based method for Incremental Semantic Segmentation, and introduces two patch-wise contrastive losses to imitate similar features and enhance feature diversity respectively, which can further improve the performance of TISS.
- 2
- 1
- 1
- 1
- 1
- DiffFAE: Advancing High-fidelity One-shot Facial Appearance Editing with Space-sensitive Customization and Semantic Preservation1
DiffFAE is presented, a one-stage and highly-efficient diffusion-based framework tailored for high-fidelity FAE, and a consistency regularization is introduced for the pipeline to enhance editing controllability by leveraging prior knowledge in the attention matrices of diffusion model.
- Multi-Forgery Detection Challenge 2022: Push the Frontier of Unconstrained and Diverse Forgery Detection1
This paper presents the solutions from the Top 3 teams of the Multi-Forgery Detection Challenge, held concurrently with the IEEE Computer Society Workshop on Biometrics at CVPR 2022 to boost the research work in the field of image forgery detection.
- IKPCA-ELM-based Intrusion Detection Method1
In order to reduce the effects of uneven sample distribution and sample attribute differences on the extraction of KPCA features, the sample attribute mean and mean square error are introduced into the Gaussian radial basis function and polynomial kernel function respectively and the two improved kernel functions are combined to construct a hybrid kernel function.
- –
- –
- –
- –
- PiT: Progressive Diffusion Transformer–
This work proposes Pseudo Shifted Window Attention (PSWA), which fundamentally mitigates global attention redundancy and proposes the Progressive Coverage Channel Allocation (PCCA) strategy that captures high-order attention without additional computational cost.
- –
- DRL: Discriminative Representation Learning with Parallel Adapters for Class Incremental Learning–
The Discriminative Representation Learning (DRL) framework is proposed, and the Decoupled Anchor Supervision (DAS) is designed to alleviate inconsistency and enable comparable feature representations across incremental stages, thereby narrowing the gap between stage-wise local optimization over a subset of data and global inference across all classes.
- –
- –
- Towards Balanced Representation Learning with Semantic Anchor Regularization–
This paper disentangle the representation regularization from the feature learning process and proposes a semantic anchor regularization (SAR) that is generated from predefined anchors that performs in the manner of visual-language alignment but is more flexible.
- –
- –
- –
- –
- –
- –
- AdapNet: Adaptive Noise-Based Network for Low-Quality Image Retrieval–
A novel setting for low-quality image retrieval is introduced, and an innovative adaptive noise-based loss function is introduced, which dynamically adjusts its focus on the gradient in accordance with image quality, thereby augmenting the learning of unknown noisy samples during training and enhancing intra-class compactness.
- –
- –
- –
- –
- –
- –
- Human Parsing via Shape Boltzmann Machine Networks–
An approach is presented that exploits Shape Boltzmann Machine networks to improve the accuracy of human body parsing, and the proposed Curve Correction method refines the final segmentation results.
- –
- –
- –
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 2026-10-12. 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.