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
- Yi Ping YangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works117 from public data
- Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in Vitro2,082
A simple semisupervised pipeline that only uses the original training set without collecting extra data, which effectively improves the discriminative ability of learned CNN embeddings and proposes the label smoothing regularization for outliers (LSRO).
- A Discriminatively Learned CNN Embedding for Person Reidentification964
This article proposes a Siamese network that simultaneously computes the identification loss and verification loss and learns a discriminative embedding and a similarity measurement at the same time, thus taking full usage of the re-ID annotations.
- Improving person re-identification by attribute and identity learning917
An attribute-person recognition (APR) network is proposed, a multi-task network which learns a re-ID embedding and at the same time predicts pedestrian attributes, and demonstrates that by learning a more discriminative representation, APR achieves competitive re-IDs performance compared with the state-of-the-art methods.
- Joint Discriminative and Generative Learning for Person Re-Identification865
This paper proposes a joint learning framework that couples re-id learning and data generation end-to-end and renders significant improvement over the baseline without using generated data, leading to the state-of-the-art performance on several benchmark datasets.
- Camera Style Adaptation for Person Re-identification652
CamStyle can serve as a data augmentation approach that smooths the camera style disparities in person re-identification and demonstrate consistent improvement in all systems regardless of the extent of over-fitting.
- Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation599
The proposed approach dynamically sets different confidence thresholds according to the prediction variance, rectifies the learning from noisy pseudo labels, and achieves significant improvements over the conventional pseudo label learning and yields competitive performance on all three benchmarks.
- Pedestrian Alignment Network for Large-scale Person Re-Identification555
This work introduces the pedestrian alignment network (PAN) which allows discriminative embedding learning pedestrian alignment without extra annotations, and observes that when the convolutional neural network learns to discriminate between different identities, the learned feature maps usually exhibit strong activations on the human body rather than the background.
- University-1652: A Multi-view Multi-source Benchmark for Drone-based Geo-localization505
It is argued that drones could serve as the third platform to deal with the geo-localization problem and propose a strong CNN baseline on this challenging dataset, named University-1652, which is the first drone-based geo- localization dataset and enables two new tasks, i.e., drone-view target localization and drone navigation.
- Each Part Matters: Local Patterns Facilitate Cross-View Geo-Localization329
It is argued that neighbor areas can be leveraged as auxiliary information, enriching discriminative clues for geo-localization, and introduced a simple and effective deep neural network, called Local Pattern Network (LPN), to take advantage of contextual information in an end-to-end manner.
- 323
- 237
- Towards Unified Text-based Person Retrieval: A Large-scale Multi-Attribute and Language Search Benchmark204
A large Multi-Attribute and Language Search dataset for text-based person retrieval, called MALS, is introduced and the feasibility of performing pre-training on both attribute recognition and image-text matching tasks in one stone is explored, achieving state-of-the-art retrieval performance via APTM on three challenging real-world benchmarks.
- VehicleNet: Learning Robust Visual Representation for Vehicle Re-Identification195
This work proposes to build a unique large-scale vehicle dataset by harnessing four public vehicle datasets, and design a simple yet effective two-stage progressive approach to learning more robust visual representation from VehicleNet.
- 169
- Macro-Micro Adversarial Network for Human Parsing165
Compared with traditional adversarial networks, the proposed Macro-Micro Adversarial Net not only enforces local and semantic consistency explicitly, but also avoids the poor convergence problem of adversarial Networks when handling high resolution images.
- Thorax disease classification with attention guided convolutional neural network136
This paper proposes to integrate the global and local cues into a three-branch attention guided convolution neural network (AG-CNN) to identify thorax diseases and demonstrates that after integrating the local cues with the global information, the average AUC scores are improved by AG-CNN.
- Unsupervised Scene Adaptation with Memory Regularization in vivo130
This paper proposes an orthogonal method, called memory regularization in vivo, to exploit the intra-domain knowledge and regularize the model training, and refers to the segmentation model itself as the memory module, and minor the discrepancy of the two classifiers to reduce the prediction inconsistency.
- PiPa: Pixel- and Patch-wise Self-supervised Learning for Domain Adaptative Semantic Segmentation96
This work revisits the current pixel contrast in semantic segmentation and proposes a unified pixel- and patch-wise self-supervised learning framework, called PiPa, for domain adaptive semantic segmentsation that facilitates intra-image pixel-wise correlations and patch -wise semantic consistency against different contexts.
- Composed Image Retrieval with Text Feedback via Multi-grained Uncertainty Regularization94
A unified learning approach to simultaneously modeling the coarse- and fine-grained retrieval by considering the multi-grained uncertainty is introduced, which prevents the model from pushing away potential candidates in the early stage, and thus improves the recall rate.
- 92
- VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimation83
VL-Uncertainty is introduced, the first uncertainty-based framework for detecting hallucinations in LVLMs that utilizes uncertainty as an intrinsic metric and significantly outperforms strong baseline methods in hallucination detection.
- Parameter-Efficient Person Re-Identification in the 3D Space80
This work projects 2D images to a 3D space and introduces a novel parameter-efficient omni-scale graph network (OG-Net) to learn the pedestrian representation directly from 3D point clouds and exploits the complementary information of 2D appearance and 3D structure.
- Dual-Path Convolutional Image-Text Embedding.77
This paper builds a convolutional network amenable for fine-tuning the visual and textual representations, where the entire network only contains four components, i.e., convolution layer, pooling layer, rectified linear unit function (ReLU), and batch normalisation.
- 75
- Progressive Local Filter Pruning for Image Retrieval Acceleration75
This article proposes a Progressive Local Filter Pruning (PLFP) method, which preserves the original model capacity while pruning and achieves superior performance to other widely-used pruning methods.
- Learning Cross-View Geo-Localization Embeddings via Dynamic Weighted Decorrelation Regularization71
This work argues that minimizing embedding redundancy is a pivotal factor in enhancing a model’s ability to discriminate diverse scene characteristics and introduces a straightforward yet effective regularization technique, termed dynamic weighted decorrelation regularization (DWDR), which significantly outperforms conventional baselines in stringent conditions.
- Towards Natural Language-Guided Drones: GeoText-1652 Benchmark with Spatial Relation Matching68
GeoText-1652 is introduced, a new natural language-guided geo-localization benchmark that extends the established University-1652 image dataset with spatial-aware text annotations, thereby establishing one-to-one correspondences between image, text, and bounding box elements.
- Multi-View Consistent Generative Adversarial Networks for 3D-aware Image Synthesis63
This work proposes Multi-View Consistent Generative Adversarial Networks (MVCGAN) for high-quality 3D-aware image synthesis with geometry constraints, and enforces the photometric consistency between pairs of views and integrates a stereo mixup mechanism into the training process.
- 63
- Decoupled and Memory-Reinforced Networks: Towards Effective Feature Learning for One-Step Person Search56
To reconcile the conflicts of multiple objectives, the standard tightly coupled pipelines are simplified and a deeply decoupled multi-task learning framework is established and a memory-reinforced mechanism is built to boost the identification feature learning.
- Adaptive Boosting for Domain Adaptation: Toward Robust Predictions in Scene Segmentation51
This paper proposes one efficient bootstrapping method, called Adaboost Student, explicitly learning complementary models during training and liberating users from empirical early stopping, which is orthogonal to most domain adaptation methods and can be combined with existing approaches to further improve the state-of-the-art performance.
- 46
- From Data Deluge to Data Curation: A Filtering-WoRA Paradigm for Efficient Text-based Person Search39
A new Filtering-WoRA paradigm is introduced, which contains a filtering algorithm to identify this crucial data subset and WoRA (Weighted Low-Rank Adaptation) learning strategy for light fine-tuning.
- Connecting Language and Vision for Natural Language-Based Vehicle Retrieval39
This work proposes to jointly train the state-of-the-art vision models with the transformer-based language model in an end-to-end manner to connect language and vision, and achieves the 1st place on the 5th AI City Challenge.
- 36
- SPG-VTON: Semantic Prediction Guidance for Multi-Pose Virtual Try-on35
The qualitative and quantitative experiments show that SPG-VTON is superior to the state-of-the-art methods and is robust to data noise, including background and accessory changes, i.e., hats and handbags, showing good scalability to the real-world scenario.
- Self-Supervised Point Cloud Representation Learning via Separating Mixed Shapes34
A new self-supervised learning method, called Mixing and Disentangling (MD), for 3D point cloud representation learning, which has improved the empirical performance on both ModelNet-40 and ShapeNet-Part datasets in terms of point cloud classification and segmentation tasks.
- Jointly Harnessing Prior Structures and Temporal Consistency for Sign Language Video Generation33
The Structure-aware Temporal Consistency Network (STCNet) is proposed, which jointly optimizes the prior structure of humans with temporal consistency to produce sign language videos to address the issue of finger motions in generated videos.
- 33
- 33
- StepNet: Spatial-temporal Part-aware Network for Isolated Sign Language Recognition32
A new framework called the Spatial-temporal Part-aware network (StepNet), based on RGB parts, made up of two modules: Part-level Spatial Modeling and Part-level Temporal Modeling, which automatically captures the appearance-based properties in the feature space without the use of any keypoint-level annotations.
- The 9th AI City Challenge29
The ninth AI City Challenge continues to advance real-world applications of computer vision and AI in transportation, industrial automation, and public safety, and saw a 17% increase in participation, with 245 teams from 15 countries registered on the evaluation server.
- UAVM '23: 2023 Workshop on UAVs in Multimedia: Capturing the World from a New Perspective27
This workshop covers various topics related to UAV multimedia, including aerial image and video processing, machine learning for UAV data analysis, UAV swarm technology, and UAV-based multimedia applications.
- Multi-view Consistent Generative Adversarial Networks for Compositional 3D-Aware Image Synthesis27
This paper proposes multi-view consistent generative adversarial networks (MVCGAN) for compositional 3D-aware image synthesis, and forms the multi-object scene generation as a “decompose and compose” process.
- Cognitive Accident Prediction in Driving Scenes: A Multimodality Benchmark27
This work proposes a Cognitive Accident Prediction (CAP) method that explicitly leverages human-inspired cognition of text description on the visual observation and the driver attention to facilitate model training and validated the superiority of CAP compared with state-of-the-art approaches.
- Open Set Adversarial Examples26
This work views open set recognition as a retrieval task and proposes a new approach, Opposite-Direction Feature Attack (ODFA), to generate adversarial examples / queries, and addresses the two limitations of closed set attack methods, ODFA directly works on the features for retrieval.
- Learnable Pillar-based Re-ranking for Image-Text Retrieval25
This paper designs a neighbor-aware graph reasoning module to flexibly exploit the relations and excavate the sparse positive items within a neighborhood, and presents a structure alignment constraint to promote crossmodal collaboration and align the asymmetric modalities.
- 25
- 25
- Transferring to Real-World Layouts: A Depth-aware Framework for Scene Adaptation24
A depth-aware framework to explicitly leverage depth estimation to mix categories and facilitate two complementary tasks, i.e., segmentation and depth learning in an end-to-end manner is proposed.
- Collaborative group: Composed image retrieval via consensus learning from noisy annotations24
The Consensus Network (Css-Net), inspired by the psychological concept that groups outperform individuals, is proposed, which comprises a consensus module with four diverse compositors, each generating distinct image-text embeddings, fostering complementary feature extraction and mitigating dependence on any single, potentially biased compositor.
- 23
- Real-World Automatic Makeup via Identity Preservation Makeup Net21
This paper introduces a new makeup model, called Identity Preservation Makeup Net (IPM-Net), which preserves not only the background but the critical patterns of the original identity, and disentangle the face images to two different information codes, i.e., identity content code and makeup style code.
- Bayesian query expansion for multi-camera person re-identification20
This paper proposes the Bayesian Query Expansion (BQE) method to generate a new query with information from the initial ranking list, and applies pooling on the features of “true matches” to get a single vector, i.e., the expanded new query.
- CAMeL: Cross-Modality Adaptive Meta-Learning for Text-Based Person Retrieval19
A domain-agnostic pretraining framework based on Cross-modality Adaptive Meta-Learning (CAMeL) to enhance the model generalization capability during pretraining to facilitate the subsequent downstream tasks and showcases robustness and scalability in handling biased synthetic images and noisy text annotations.
- Ctrl-U: Robust Conditional Image Generation via Uncertainty-aware Reward Modeling19
This paper proposes an uncertainty-aware reward modeling, called Ctrl-U, including uncertainty estimation and uncertainty-aware regularization, designed to reduce the adverse effects of imprecise feedback from the reward model.
- U-Turn: Crafting Adversarial Queries with Opposite-Direction Features18
The main idea underpinning ODFA is to impel the original image feature to the opposite direction, similar to a U-turn, which shows that the adversarial queries generated by ODFA cause true matches no longer to be seen at the top ranks, and the attack success rate is consistently higher than classifier attack methods.
- Beyond Walking: A Large-Scale Image-Text Benchmark for Text-Based Person Anomaly Search17
A large-scale imagetext Pedestrian Anomaly Behavior (PAB) benchmark is constructed, and a cross-modal pose- aware framework is introduced, which integrates human pose patterns with identity-based hard negative pair sampling, to validate the potential of PAB.
- Scale Up Composed Image Retrieval Learning via Modification Text Generation14
This paper proposes to synthesize the training triplets to augment the training resource for the CIR problem and develops a two-hop alignment strategy to incrementally close the semantic gap between the multimodal pair and the target image.
- Progressive Text-to-3D Generation for Automatic 3D Prototyping13
This work presents a novel Multi-Scale Triplane Network (MTN) architecture coupled with a tailored progressive learning strategy that systematically guides the network to shift its attention from prominent coarse-grained structures to intricate fine-grained patterns.
- The 3rd Workshop on UAVs in Multimedia: Capturing the World from a New Perspective13
This workshop will cover various topics related to UAV multimedia, including aerial image and video processing, machine learning for UAV data analysis, UAV swarm technology, and UAV-based multimedia applications.
- The 2nd Workshop on UAVs in Multimedia: Capturing the World from a New Perspective13
This workshop will cover various topics related to UAV multimedia, including aerial image and video processing, machine learning for UAV data analysis, UAV swarm technology, and UAV-based multimedia applications.
- 3D Magic Mirror: clothing reconstruction from a single image via a causal perspective12
This paper proposes a causality-aware self-supervised learning method to adaptively reconstruct 3D non-rigid objects from 2D images without 3D annotations and introduces an explainable structural causal map (SCM) to build the model.
- Deep Multimodal Learning for Information Retrieval11
This workshop aims to provide a platform for discussion about multi-modal IR among scholars, practitioners, and other interested parties and show the great potential shown by multimodal-empowered IR.
- Person Re-identification in the 3D Space.10
This work project 2D images to a 3D space and introduce a novel Omni-scale Graph Network (OG-Net) to learn the representation from sparse 3D points, which can learn a new type of deep re-id feature free from noisy variants, such as scale and viewpoint.
- Harnessing Uncertainty-Aware Bounding Boxes for Unsupervised 3D Object Detection9
A new uncertainty-aware framework for unsupervised 3D object detection, dubbed UA3D, which incorporates an extra auxiliary detection branch alongside the original primary detector and adaptively adjusts the weight of every 3D bbox coordinate via uncertainty regularization, refining the training process on pseudo bboxes.
- 9
- 8
- Active Discovering New Slots for Task-Oriented Conversation8
A general new slot discovery task is formulated in an information extraction fashion and incorporate it into an active learning framework to realize human-in-the-loop learning and proposes a bi-criteria selection scheme which incorporates two major strategies, namely, uncertainty-based and diversity-based sampling to efficiently identify terms of interest.
- 8
- Generating Attribution Reports for Manipulated Facial Images: A Dataset and Baseline7
This work introduces Forgery Attribution Report Generation, a new multimodal task that jointly localizes forged regions and generates natural language explanations grounded in the editing process and proposes ForgeryTalker, a unified end-to-end framework that integrates vision and language via a shared encoder and dual decoders for mask and text generation, enabling coherent cross-modal reasoning.
- 7
- Instilling Multi-round Thinking to Text-guided Image Generation7
A new self-supervised regularization, multi-round regularization, which is compatible with existing methods is introduced, and builds upon the observation that the modification order generally should not affect the final result.
- 7
- Negative sequential pattern mining7
Three innovative algorithms are proposed to mine NSP efficiently and extensive experiments about the three algorithms on either synthetic or real-world datasets show that the proposed methods can find NSP efficient.
- CLIP-SR: Collaborative Linguistic and Image Processing for Super-Resolution6
This work proposes a multi-modal semantic enhancement framework that integrates textual semantics with visual features, effectively mitigating semantic mismatches and detail losses in highly degraded low-resolution images and offers a degree of super-resolution editability, allowing for controlled enhancements while preserving semantic consistency.
- Uncertainty-o: One Model-agnostic Framework for Unveiling Uncertainty in Large Multimodal Models6
Uncertainty-o is introduced, a model-agnostic framework designed to reveal uncertainty in LMMs regardless of their modalities, architectures, or capabilities, and derive the formulation of multimodal semantic uncertainty, which enables quantifying uncertainty from multimodal responses.
- UniAD: Integrating Geometric and Semantic Cues for Unified Anomaly Detection6
UnicAD is proposed, a novel dual-branch teacher-student framework that achieves unified anomaly detection through synergistic integration of complementary expertise from heterogeneous vision models without requirements of extra manual annotations.
- Towards Scalable Video Anomaly Retrieval: A Synthetic Video-Text Benchmark5
SVTA (Synthetic Video-Text Anomaly benchmark), the first large-scale dataset for cross-modal anomaly retrieval, leveraging generative models to overcome data availability challenges is introduced, revealing SVTA's challenging nature and its effectiveness in evaluating a robust cross-modal retrieval method.
- SEED: A Benchmark Dataset for Sequential Facial Attribute Editing with Diffusion Models5
This work introduces SEED, a large-scale Sequentially Edited facE Dataset constructed via state-of-the-art diffusion models, and proposes FAITH, a frequency-aware transformer-based model that incorporates high-frequency cues to enhance sensitivity to subtle sequential changes.
- When Words Smile: Generating Diverse Emotional Facial Expressions from Text4
This work introduces an end-to-end text-to-expression model that explicitly focuses on emotional dynamics and learns expressive facial variations in a continuous latent space and generates expressions that are diverse, fluid, and emotionally coherent.
- 4
- High Fidelity Makeup via 2D and 3D Identity Preservation Net4
The High Fidelity Makeup via two-dimensional and 3D Identity Preservation Network (IP23-Net) is proposed, to the best of the knowledge, a novel framework that leverages facial geometry information to generate more realistic results.
- MORE'24 Multimedia Object Re-ID: Advancements, Challenges, and Opportunities4
A spectrum of topics related to object re-id are covered, including but not limited to deep metric learning, multi-view data generation, video-based object re-id, cross-domain object re-id and real-world applications.
- 4
- 4
- EQ-TAA: Equivariant Traffic Accident Anticipation via Diffusion-Based Accident Video Synthesis3
An Attentive Video Diffusion model is proposed that synthesizes additional accident video clips by generating the causal part in dashcam videos, i.e., from normal clips to accident clips, while preserving the style and content of frames for TAA after video generation.
- 3
- Harnessing weak pair uncertainty for text-based person search2
This paper introduces an uncertainty-aware method to explicitly estimate image-text pair uncertainty, and incorporate the uncertainty into the optimization procedure in a smooth manner, and proposes the uncertainty regularization to adaptively adjust loss weight.
- 2
- Domain-Agnostic Neural Oil Painting via Normalization Affine Test-Time Adaptation2
A domain-agnostic neural painting (DANP) framework that aligns model to the test domain, and focuses on updating affine parameters of normalization layers efficiently, while keeping other parameters frozen, to stabilize adaptation.
- 2
- 2
- 2
- Every Painting Awakened: A Training-free Framework for Painting-to-Animation Generation2
By achieving enhanced dynamic effects without requiring any model training or learnable parameters, this framework enables plug-and-play integration with existing I2V methods, making it an ideal solution for animating real-world paintings.
- Multimedia Content Understanding in Harsh Environments2
This tutorial introduces some multimedia enhancement methods, presents recent advances in 2D and 3D visual scene understanding, shows strategies to estimate the prediction uncertainty, provides a brief summary, and shows some typical applications.
- Event-based Lip Reading with Triplane Fusion Network1
This paper proposes an event-specified framework named the Triplane Fusion Network (TF-Net) to harness lip movements through analysis from three distinct but complementary views, and observes that these three views complement each other, further enhancing the learning of the event-specified distribution.
- 1
- Look, Compare and Draw: Differential Query Transformer for Automatic Oil Painting1
This work proposes the Differential Query Transformer (DQ-Transformer), a new architecture that leverages differentially derived image representations enriched with positional encoding to guide the stroke prediction process, and incorporates adversarial training into the framework, enhancing the accuracy of stroke prediction and thereby improving the overall realism and fidelity of the synthesized paintings.
- 1
- 1
- Actively Discovering New Slots for Task-oriented Conversation1
This work formulate a general new slot discovery task in an information extraction fashion and incorporate it into an active learning framework to realize human-in-the-loop learning and proposes a bi-criteria selection scheme which incorporates two major strategies, namely, uncertainty-based sampling and diversity- based sampling to efficiently identify terms of interest.
- TIGeR: Text-Instructed Generation and Refinement for Template-Free Hand-Object Interaction–
A new Text-Instructed Generation and Refinement (TIGeR) framework, harnessing the power of intuitive text-driven priors to steer the object shape refinement and pose estimation, which shows robustness to occlusion, while maintaining compatibility with heterogeneous prior sources.
- –
- Artificial intelligence in orbital diseases: Opportunities, challenges, and future horizons–
The current relationship between AI and orbital diseases and a three-tiered roadmap comprising OrbitBank, OrbitFM, and OrbitAgent are explored, progressing from multicenter data infrastructure and privacy-preserving collaboration to multimodal foundation models and, ultimately, clinician-supervised agentic systems for more standardized and precise orbital care.
- SynthIR: The First Workshop on Synthetic Content in Information Retrieval Ecosystems–
This workshop aims to provide a forum to analyze implications of AI-generated content implications with an emphasis on reflection, evaluation, and human-centered system design, and to foster community-driven discussion that may inform future evaluation efforts, including potential shared tasks or tracks in venues such as TREC, CLEF, FIRE, or NTCIR.
- –
- –
- MORE'25 Multimedia Object Re-ID: Advancements, Challenges, and Opportunities–
A spectrum of topics related to object re-id are covered, including but not limited to deep metric learning, multi-view data generation, video-based object re-id, cross-domain object re-id, and real-world applications.
- –
- Introduction to the Special Issue on Deep Multimodal Generation and Retrieval–
This editorial introduces the Special Issue on Deep Multimodal Generation and Retrieval, providing a platform for researchers to share insights and advancements in this rapidly evolving domain and discusses ongoing challenges and future opportunities in this rapidly evolving domain.
- The 2nd International Workshop on Deep Multi-modal Generation and Retrieval–
This workshop aims to encourage more explorations in Deep Multimodal Generation and Retrieval, providing a platform for researchers to share insights and advancements in this rapidly evolving domain.
- –
- –
- –
- Subband-based Generative Adversarial Network for Non-parallel Many-to-many Voice Conversion–
Inspired by the inherent structure of mel-spectrogram, a new voice conversion framework, i.e., Subband-based Generative Adversarial Network for Voice Conversion (SGAN-VC), which converts each subband content of the source speech separately by explicitly utilizing the spatial characteristics between different subbands.
- –
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.