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
- Ming–Hsuan YangSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works117 from public data
- The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking1,076
This work constructs a new UAV benchmark focusing on complex scenarios with new level challenges, and is the first time to explore issues in unconstrained scenes comprehensively in UAV based real scenes.
- The Visual Object Tracking VOT2016 Challenge Results776
The Visual Object Tracking challenge VOT2016 goes beyond its predecessors by introducing a new semi-automatic ground truth bounding box annotation methodology and extending the evaluation system with the no-reset experiment.
- 705
- The Visual Object Tracking VOT2014 Challenge Results588
The Visual Object Tracking challenge 2014, VOT2014, aims at comparing short-term single-object visual trackers that do not apply pre-learned models of object appearance and introduces a new dataset with full annotation of targets by rotated bounding boxes and per-frame attribute.
- REVERIE: Remote Embodied Visual Referring Expression in Real Indoor Environments551
A dataset of varied and complex robot tasks, described in natural language, in terms of objects visible in a large set of real images, and a novel Interactive Navigator-Pointer model is proposed that provides a strong baseline on the task.
- VLN↻BERT: A Recurrent Vision-and-Language BERT for Navigation503
A recurrent BERT model that is time-aware for use in VLN is proposed that can replace more complex encoder-decoder models to achieve state-of-the-art results and can generalised to other transformer-based architectures.
- AerialVLN: Vision-and-Language Navigation for UAVs208
A new task named AerialVLN is proposed, which is UAV-based and towards outdoor environments, and an extended baseline model based on the widely-used cross-modal-alignment (CMA) navigation methods are proposed.
- Exploiting Completeness and Uncertainty of Pseudo Labels for Weakly Supervised Video Anomaly Detection143
This work designs a multi-head classification module (each head serves as a classifier) with a diversity loss to maximize the distribution differences of predicted pseudo labels across heads and devise an iterative uncertainty pseudo label refinement strategy, which improves not only the initial pseudo labels but also the updated ones obtained by the desired classifier in the second stage.
- Object-and-Action Aware Model for Visual Language Navigation142
An Object-and-Action Aware Model (OAAM) is proposed that processes these two different forms of natural language based instruction separately and enables each process to match object-centered/action-centered instruction to their own counterpart visual perception/action orientation flexibly.
- 122
- Hierarchical Modular Network for Video Captioning111
A hierarchical modular network to bridge video representations and linguistic semantics from three levels before generating captions is proposed and shows that the proposed method performs favorably against the state-of-the-art models on the two widely-used benchmarks.
- HOP: History-and-Order Aware Pretraining for Vision-and-Language Navigation108
This work proposes a novel history-and-order aware pre-training paradigm (HOP) with VLN-specific objectives that exploit the past observations and support future action prediction to promote the learning of spatio-temporal visual-textual correspondence and the agent's capability of decision making.
- Neighbor-view Enhanced Model for Vision and Language Navigation99
This work proposes a multi-module Neighbor-View Enhanced Model (NvEM) to adaptively incorporate visual contexts from neighbor views for better textual-visual matching of existing methods.
- The Road to Know-Where: An Object-and-Room Informed Sequential BERT for Indoor Vision-Language Navigation95
This work proposes an object-informed sequential BERT to encode visual perceptions and linguistic instructions at the same fine-grained level, namely objects and words, and enables the visual-textual clues to be interpreted in light of the temporal context, which is crucial to multi-round VLN tasks.
- VisDrone-SOT2018: The Vision Meets Drone Single-Object Tracking Challenge Results87
The evaluation protocol of the VisDrone-SOT2018 challenge is presented and the results of a comparison of 22 trackers on the benchmark dataset are presented, which are publicly available on the challenge website.
- 85
- 80
- From Speaker to Dubber: Movie Dubbing with Prosody and Duration Consistency Learning71
This work proposes a two-stage dubbing method that allows the model to first learn pronunciation knowledge before practicing it in movie dubbing, and outperforms several state-of-the-art methods on two primary benchmarks.
- Multi-Attention Network for Compressed Video Referring Object Segmentation69
A multi-attention network which consists of dual-path dual-att attention module and a query-based cross-modal Transformer module designed to extract effective representation from compressed data in three modalities, i.e., I-frame, Motion Vector and Residual is proposed.
- March in Chat: Interactive Prompting for Remote Embodied Referring Expression65
A March-in-Chat (MiC) model that can talk to the LLM on the fly and plan dynamically based on a newly proposed Room-and-Object Aware Scene Perceiver (ROASP) and outperforms the previous state-of-the-art by large margins by SPL and RGSPL metrics on the REVERIE benchmark.
- 61
- Diagnosing Vision-and-Language Navigation: What Really Matters59
This work conducts a series of diagnostic experiments to unveil agents’ focus during navigation and shows that indoor navigation agents refer to both object and direction tokens when making decisions, and Transformer-based agents acquire a better cross-modal understanding of objects and display strong numerical reasoning ability than non-Transformer- based agents.
- Learning to Dub Movies via Hierarchical Prosody Models54
This work proposes a novel movie dubbing architecture to tackle these problems via hierarchical prosody modeling, which bridges the visual information to corresponding speech prosody from three aspects: lip, face, and scene.
- 53
- Point-to-Set Distance Metric Learning on Deep Representations for Visual Tracking52
This paper proposes to integrate the point-to-set/ image- to-imageSet distance metric learning (DML) into visual tracking tasks and take full advantage of all the training samples when determining the best target candidate.
- Plant identification based on very deep convolutional neural networks50
A deep learning method is employed to extract discriminative features from plant images along with a linear SVM for plant identification and the remarkable performance of the very deep neural network compared to the hand-crafted features is demonstrated.
- Learning Attribute-Specific Representations for Visual Tracking48
It is shown that attribute information in the context facilitates training an effective classifier for visual tracking, and an attribute-based CNN with multiple branches is designed, where each branch is responsible for classifying the target under a specific attribute.
- V2C: Visual Voice Cloning43
This work designs a set of evaluation metrics, named MCD-DTW-SL, which help eval-uate the similarity between ground-truth speeches and the synthesised ones, and proposes a new task, V2C, which seeks to convert a para-graph of text to a speech with both desired voice speci-fied by a reference audio and desired emotion specified by areference video.
- 43
- Decomposing Disease Descriptions for Enhanced Pathology Detection: A Multi-Aspect Vision-Language Pre-Training Framework42
This paper introduces a novel VLP framework designed to dissect disease descriptions into their fundamental aspects, leveraging prior knowledge about the visual manifestations of pathologies, and presents a dual-head Transformer tailored to process known and unknown diseases, optimizing the comprehensive detection efficacy.
- Siamese Local and Global Networks for Robust Face Tracking41
A face tracking method based on Siamese CNNs, which takes advantages of powerful representations of hierarchical CNN features learned from massive face images, which demonstrates the effectiveness of the method in comparison to several state-of-the-art visual tracking methods.
- Generating Content for HDR Deghosting from Frequency View40
The Low-Frequency aware Diffusion (LF-Diff) model is proposed, implementing the DMs in a highly compacted latent space and integrating it into a regression-based model to enhance the details of reconstructed images.
- BEVBert: Multimodal Map Pre-training for Language-guided Navigation40
This work introduces hybrid topo-metric maps into VLN, where a topological map is used for long-term planning and a metric map for short-term reasoning, and designs a pre-training framework via the hybrid map to learn language-informed map representations, which enhances cross-modal grounding and facilitates the language-guided navigation goal.
- 38
- 38
- Robust visual tracking via scale-and-state-awareness35
This work proposes to adaptively utilize the level set segmentation and bounding box regression techniques to obtain a tight enclosing box, and design a CNN to recognize whether the target is occluded.
- EmoDubber: Towards High Quality and Emotion Controllable Movie Dubbing34
EmoDubber is proposed, an emotion-controllable dubbing architecture that allows users to specify emotion type and emotional intensity while satisfying high-quality lip sync and pronunciation.
- 33
- 30
- 27
- Augmented Commonsense Knowledge for Remote Object Grounding23
This work proposes an augmented commonsense knowledge model (ACK), which consists of knowledge graph-aware cross-modal and concept aggregation modules to enhance visual representation and visual-textual data alignment by integrating visible objects, commonsense knowledge, and concept history, which includes object and knowledge temporal information.
- Progressive Multi-Resolution Loss for Crowd Counting23
This paper proposes to predict the density map at one resolution but measure its quality via a derived log-formed loss at multiple resolutions, and mathematically proves that the traditional single-resolution L2 loss is superior to a single-resolution L2 loss.
- 21
- Presentation Attack Detection: A Systematic Literature Review20
This systematic literature review identifies and categorises the state-of-the-art approaches in each domain to cover the challenges and solutions in each domain to cover the challenges and solutions in a single place.
- Light fixed-time control for cluster synchronization of complex networks20
A light fixed-time controller is proposed with a compact formulation by revisiting continuous control paradigm, and sufficient conditions are obtained by Lyapunov stability theorem to guarantee the nodes in each cluster to desired states in a settling time.
- RETTA: Retrieval-enhanced test-time adaptation for zero-shot video captioning19
A novel zero-shot video captioning framework named Retrieval-Enhanced Test-Time Adaptation (RETTA), which takes advantage of existing pretrained large-scale vision and language models to directly generate captions with test-time adaptation.
- 19
- R-GAN: Exploring Human-like Way for Reasonable Text-to-Image Synthesis via Generative Adversarial Networks19
This work proposes R-GAN, which can generate reasonable images according to the given text in a human-like way, and introduces a coarse-to-fine shape generator based on a shape knowledge base.
- 19
- Release the Power of Online-Training for Robust Visual Tracking18
This paper proposes to improve the tracking accuracy via online training by squeezing redundant training data by analyzing the dataset distribution in low-level feature space and design statistic-based losses to increase the inter-class distance while decreasing the intra-class variance of high-level semantic features.
- Consistency-Aware Anchor Pyramid Network for Crowd Localization17
A Consistency-Aware Anchor Pyramid Network (CAAPN) comprising an Adaptive Anchor Generator (AAG) and a Localizer with Augmented Matching (LAM) that achieves favorable performance against state-of-the-art approaches on five challenging datasets.
- 17
- 17
- 16
- Visual and Semantic Prompt Collaboration for Generalized Zero-Shot Learning15
A visual prompt to integrate the visual information for discriminative feature learning and a semantic prompt to integrate the semantic formation for visual-semantic alignment is designed and can obtain discriminative semantic-related features for generalized zero-shot image recognition.
- 14
- Overwater Image Dehazing via Cycle-Consistent Generative Adversarial Network14
This work collects the first overwater image dehazing dataset and proposes a Generative Adversial Network (GAN)-based method called OverWater Image Dehazing GAN (OWI-DehazeGAN), composed of an encoder–decoder framework, supervised by a forward-backward translation consistency loss for self-supervision and a perceptual loss for content preservation.
- 13
- Prosody-Enhanced Acoustic Pre-training and Acoustic-Disentangled Prosody Adapting for Movie Dubbing13
This work proposes an acoustic-prosody disentangled two-stage method to achieve high-quality dubbing generation with precise prosody alignment and incorporates an in-domain emotion analysis module to reduce the impact of visual domain shifts across different movies, thereby enhancing emotion-prosody alignment.
- A Unified Object Counting Network With Object Occupation Prior13
This paper builds the first evolving object counting dataset and proposes a unified object counting network that can efficiently and effectively adapt to new classes while maintaining good performance on already-seen data without large-scale retraining.
- High Performance Gesture Recognition via Effective and Efficient Temporal Modeling13
A simple and efficient architectural unit, Multi-Kernel Temporal Block (MKTB), that models the multi-scale temporal responses by explicitly applying different temporal kernels and a Global Refinement Block (GRB), which is an attention module for shaping the global temporal features based on the cross-channel similarity.
- 11
- CALM: An Enhanced Encoding and Confidence Evaluating Framework for Trustworthy Multi-view Learning11
The advantage of CALM lies in its ability to evaluate the quality of each view, reducing the influence of low-quality views on the multi-view fusion process and ultimately leading to improved classification performance and confidence evaluation.
- 10
- FlowDubber: Movie Dubbing with LLM-based Semantic-aware Learning and Flow Matching based Voice Enhancing9
A novel dubbing architecture based on Large Language Model (LLM) and Conditional Flow Matching (CFM), named FlowDubber, which achieves high-quality audio-visual sync and pronunciation by incorporating a large speech language model with dual contrastive alignment while improving acoustic quality via Flow-based Voice Enhancing (FVE).
- 9
- 9
- Weakly Supervised Video Individual Counting9
A weakly supervised VIC task is introduced, wherein trajectory labels are not provided, and two types of labels are provided to indicate traffic entering the field of view (inflow) and leaving the field view (outflow), and results show that the baseline weakly supervised method outperforms supervised methods.
- 9
- 8
- Tracking the Unstable: Appearance-Guided Motion Modeling for Robust Multi-Object Tracking in UAV-Captured Videos8
This work proposed AMOT, which jointly exploits appearance and motion cues through two key components: an Appearance-Motion Consistency (AMC) matrix and a Motion-aware Track Continuation (MTC) module, enabling more reliable and context-aware identity association.
- Seeing the Trees for the Forest: Rethinking Weakly-Supervised Medical Visual Grounding7
Simple, yet effective Disease-Aware Prompting (DAP) process is introduced, which uses the explainability map of a VLM to identify the appropriate image features and amplifies disease-relevant regions while suppressing background interference.
- Incomplete Multi-View Multi-Label Classification via Diffusion-Guided Redundancy Removal7
This paper makes the first attempt to take advantage of diffusion models to address the missing view problem and design a strategy to identify and remove redundant views.
- Eye-Guided Multimodal Fusion: Toward an Adaptive Learning Framework Using Explainable Artificial Intelligence7
This work proposes an Eye-Gaze Guided Multimodal Fusion framework that leverages expert eye-tracking data to enhance learning and decision-making in medical image interpretation and offers a promising pathway toward intelligent, human-centered AI systems that support both diagnostic accuracy and medical training.
- RERERE: Remote Embodied Referring Expressions in Real indoor Environments.7
This work proposes the first Remote Embodied Referring Expression dataset of natural language references to remote objects in real images, which represents a complex challenge, but one that closely reflects one of the core visual problems in robotics.
- InstructDubber: Instruction-based Alignment for Zero-shot Movie Dubbing6
This work proposes InstructDubber, a novel instruction-based alignment dubbing method for both robust in-domain and zero-shot movie dubbing that outperforms state-of-the-art approaches across both in-domain and zero-shot scenarios.
- Dynamic Erasing Network Based on Multi-Scale Temporal Features for Weakly Supervised Video Anomaly Detection5
A Dynamic Erasing Network (DE-Net) is proposed for weakly supervised video anomaly detection, which learns multi-scale temporal features and designs a dynamic erasing strategy, which dynamically assesses the completeness of the detected anomalies and erases prominent abnormal segments in order to encourage the model to discover gentle abnormal segment in a video.
- Image editing with varying intensities of processing4
The hypothesis is that the part of parameters that correspond to the intensity of processing of the network are implicitly embedded in the whole parameter set and can be explicitly extracted and utilized, even with limited supervision, in image editing with Varying Intensities of Processing.
- STGS: Spatio-temporal Graph Sparsification Using Reinforcement Learning3
STGS (Spatio-Temporal Graph Sparsification), a reinforcement learning-based framework for sparsifying spatio-temporal graphs, is introduced, demonstrating that STGS outperforms existing methods in both structural preservation and downstream forecasting tasks.
- Trffc: Efficient Traffic Forecasting through Adaptive Spatio-Temporal Graph Reduction3
Tffc is introduced, a reinforcement learning-based framework to sparsify temporal traffic graphs, enhancing both the scalability of forecasting and the recognition of emerging patterns, and significantly boosts the performance of traffic forecasting models.
- 3
- SDVPT: Semantic-Driven Visual Prompt Tuning for Open-world Object Counting3
This work proposes a plug-and-play Semantic-Driven Visual Prompt Tuning framework (SDVPT) that transfers knowledge from the training set to unseen categories with minimal overhead in parameters and inference time.
- 3
- 3
- CausalMVC: Causal Content-Style Representation Learning for Deep Multi-View Clustering3
CausalMVC incorporates causal content-style disentanglement via a dual differential content-style network for separation of semantic information from noise and introduces causal content consistency that aligns semantic content from both intra-view and cross-view perspectives.
- A Generative Approach for Comprehensive Financial Event Extraction at the Document Level3
This paper proposes a generative solution for document-level event extraction, which is more challenging due to the presence of scattered arguments and multiple events, and introduces an encoding scheme to capture entity-to-document level information and a decoding scheme that makes the generative process aware of all relevant contexts.
- 2
- Exploring Primitive Visual Measurement Understanding and the Role of Output Format in Learning in Vision-Language Models2
The findings reveal that coherent sentence-based outputs outperform tuple formats, particularly in OD scenarios with large domain gaps, and it is demonstrated that scaling numeric tokens during loss computation enhances numerical approximation capabilities, further improving performance on spatial and measurement tasks.
- 2
- Adapter-Enhanced Semantic Prompting for Continual Learning2
A novel lightweight CL framework, Adapter-Enhanced Semantic Prompting (AESP), which integrates prompt tuning and adapter techniques to enhance the generalization ability of visual features and utilize adapters to efficiently fuse the semantic information, aiming to learn more adaptive features for the continual learning task.
- 2
- Link prediction on multi-relational graphs from an influence propagation perspective1
The Influential Graph Neural Predictor, referred to as IGNP, a link prediction framework guided by influence propagation, is proposed, which outperforms strong baselines by a large margin on the widely used and real-world datasets.
- A Survey on Improving Human Robot Collaboration through Vision-and-Language Navigation1
It is advocated that the future VLN systems should support proactive clarification, real-time feedback, and contextual reasoning through advanced natural language understanding (NLU) techniques, which are essential for scalable, efficient multi-robot collaboration.
- 1
- 1
- 1
- 1
- A Continuous Semantic Embedding Method for Video Compact Representation1
A novel continuous video semantic embedding model is designed to learn the actual distribution of video words, integrated with a well-designed discriminative negative sampling approach, which helps emphasize the convincing clips in the embedding while weakening the influence of the confusing ones.
- –
- Dual Adaptive Visual-Semantic Prompt Collaboration for Generalized Zero-Shot Learning–
The Dual Adaptive Visual-Semantic Prompt Collaboration Network (VSPCN+), a novel framework that utilizes prompt-tuning for effective feature adaptation, is proposed, which introduces a dual-prompt mechanism comprising both visual and semantic prompts.
- ImVision: Adapting Pretrained Vision Models for Time-Series Imputation–
ImVision is introduced, a cross-modal framework that reframes time-series imputation as image inpainting and outperforms state-of-the-art models in standardized MAE on 8 out of 10 settings, especially at high missing rates.
- –
- –
- DGRNet: Disagreement-Guided Refinement for Uncertainty-Aware Brain Tumor Segmentation–
The Disagreement-Guided Refinement Network (DGRNet) is proposed, a novel framework that addresses both limitations through multi-view disagreement-based uncertainty estimation and text-conditioned refinement and introduces a diversity-preserving training strategy that combines pairwise similarity penalties and gradient isolation to prevent view collapse.
- –
- Auditing Artificial Intelligence Systems: A Survey of Current Frameworks, Principles and Approaches–
A systematic review of the literature on methodologies, frameworks, and techniques for auditing AI systems, focusing on legal and ethical considerations and compliance with regulations, reveals gaps in current auditing practices and highlights the importance of incorporating AI value chain stages and AI maturity levels into auditing frameworks.
- –
- WPF-Mamba: wavelet-based progressive multispectral fusion mamba for fine-grained microorganism detection–
This work proposes a multispectral framework for fine-grained microorganisms detection named WPF-Mamba (Wavelet-Progressive Fusion Mamba), which incorporates a wavelet-based fusion and progressive feature refinement strategy that presents a practical paradigm for multispectral fine-grained microorganisms analysis, which in turn contributes to the development of reliable, scalable environmental monitoring systems.
- ProgFormer: Hierarchical Voxel Diffusion Transformer for Longitudinal Brain MRI Prediction–
ProgFormer, a hierarchical voxel-space Diffusion Transformer for longitudinal brain MRI prediction, is proposed, a hierarchical voxel-space Diffusion Transformer for longitudinal brain MRI prediction that models overall brain structure and longitudinal context.
- –
- –
- –
- –
- Teaching Prompts to Coordinate: Hierarchical Layer-Grouped Prompt Tuning for Continual Learning–
A novel hierarchical layer-grouped prompt tuning method for continual learning that improves model stability in two ways: (i) Layers in the same group share roughly the same prompts, which are adjusted by position encoding, and (ii) It utilizes a single task-specific root prompt to learn to generate sub-prompts for each layer group.
- –
- Towards Subject-Oriented Video Captioning via User-Specified Targets–
This work proposes a new video captioning task, subject-oriented video captioning, which allows users to specify the describing target via a bounding box, and constructs two subject-oriented video captioning datasets based on two widely used video captioning datasets: MSVD and MSRVTT.
- –
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.