About
I am a Ph.D. student at Jiangnan University, supervised by Prof. Xiao-Jun Wu. I am currently a visiting Ph.D. researcher at the Centre for Vision, Speech and Signal Processing (CVSSP), University of Surrey, supervised by Prof. Josef Kittler.
My research is on video understanding, with a focus on human action recognition and multimodal understanding of human-human interaction: learning from few or zero labelled examples, modelling skeletons, localizing actions in long, untrimmed videos, and recognizing events in social interactions from video, audio and language.
I also enjoy turning research into working systems. With my teammates I have placed in the top three at challenges held with CVPR, ICCV, ECCV and ICME.
- Action recognition
- Video understanding
- Human interaction understanding
- Multi-modal learning
- Few- & zero-shot learning
- Skeleton-based modelling
- Temporal action localization
News
- Won 2nd place in Multimodal Exocentric Event Recognition and 3rd place in Multimodal Exocentric Causal Event Grounding at the ChaLearn UDIVA-HHOI Challenge, held with the CONTEXTUS workshop at ECCV 2026.
- Won 2nd place in the Temporal Sound Localisation track of the Perception Test Challenge at ECCV 2024.
- Two papers accepted to ECCV 2024: EMP-Net for few-shot action recognition and C2C for zero-shot compositional action recognition.
- Won 2nd place in two tracks of the ICME 2024 MMVRAC Challenge, skeleton-based action recognition and person re-identification. Our solution paper appears at ICME Workshops 2024.
- Won 3rd place and the Most Novel Submission award in the Temporal Sound Localisation track of the Perception Test Challenge at ICCV 2023.
Show earlier news
- Won 2nd place in the Naturalistic Driving Action Recognition track of the AI City Challenge at CVPR 2023.
- Our paper on naturalistic driving action localization is accepted to the CVPR 2023 Workshops.
Publications
-
ECCV 2024
Efficient Few-Shot Action Recognition via Multi-Level Post-Reasoning
European Conference on Computer Vision (ECCV), 2024
Paper CodeBibTeX
@inproceedings{wu2024efficient, title = {Efficient Few-Shot Action Recognition via Multi-Level Post-Reasoning}, author = {Wu, Cong and Wu, Xiao-Jun and Li, Linze and Xu, Tianyang and Feng, Zhenhua and Kittler, Josef}, booktitle = {European Conference on Computer Vision (ECCV)}, year = {2024}, doi = {10.1007/978-3-031-72646-0_3} } -
ECCV 2024
C2C: Component-to-Composition Learning for Zero-Shot Compositional Action Recognition
European Conference on Computer Vision (ECCV), 2024 Oral
arXiv CodeBibTeX
@inproceedings{li2024c2c, title = {C2C: Component-to-Composition Learning for Zero-Shot Compositional Action Recognition}, author = {Li, Rongchang and Feng, Zhenhua and Xu, Tianyang and Li, Linze and Wu, Xiao-Jun and Awais, Muhammad and Atito, Sara and Kittler, Josef}, booktitle = {European Conference on Computer Vision (ECCV)}, year = {2024}, doi = {10.1007/978-3-031-72920-1_21} } -
ICMEW 2024
A Hybrid Multi-Perspective Complementary Model for Human Skeleton-Based Action Recognition
IEEE International Conference on Multimedia and Expo Workshops (ICMEW), 2024
2nd place, ICME 2024 MMVRAC skeleton-based action recognition
Paper CodeBibTeX
@inproceedings{li2024hybrid, title = {A Hybrid Multi-Perspective Complementary Model for Human Skeleton-Based Action Recognition}, author = {Li, Linze and Zhou, Youwei and Hu, Jiannan and Wu, Cong and Xu, Tianyang and Wu, Xiao-Jun}, booktitle = {2024 IEEE International Conference on Multimedia and Expo Workshops (ICMEW)}, pages = {1--6}, year = {2024}, organization = {IEEE} } -
arXiv 2024
TENet: Targetness Entanglement Incorporating with Multi-Scale Pooling and Mutually-Guided Fusion for RGB-E Object Tracking
arXiv preprint, 2024
arXiv CodeBibTeX
@article{shao2024tenet, title = {TENet: Targetness Entanglement Incorporating with Multi-Scale Pooling and Mutually-Guided Fusion for RGB-E Object Tracking}, author = {Shao, Pengcheng and Xu, Tianyang and Tang, Zhangyong and Li, Linze and Wu, Xiao-Jun and Kittler, Josef}, journal = {arXiv preprint arXiv:2405.05004}, year = {2024} } -
CVPRW 2023
Action Probability Calibration for Efficient Naturalistic Driving Action Localization
IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2023
2nd place, AI City Challenge 2023 Track 3
Paper CodeBibTeX
@inproceedings{li2023action, title = {Action Probability Calibration for Efficient Naturalistic Driving Action Localization}, author = {Li, Rongchang and Wu, Cong and Li, Linze and Shen, Zhongwei and Xu, Tianyang and Wu, Xiao-Jun and Li, Xi and Lu, Jiwen and Kittler, Josef}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, year = {2023} }
Honors & Competitions
Projects
-
multi-modal-tsm
Official baseline for Track 3, Multi-Modal Action Recognition, of the ICPR 2024 Multi-Modal Visual Pattern Recognition Challenge.
-
UAV-SAR
An ensemble of three skeleton models for action recognition on drone footage, 2nd place at ICME 2024 MMVRAC.
-
UAV_ReID
Person re-identification on drone footage built on AGW, 2nd place at ICME 2024 MMVRAC.
-
AICity2023 Driving Action
Driving action localization with probability calibration, 2nd place at the AI City Challenge 2023.
-
HuggingFace-Manager
A desktop app for searching, downloading and managing Hugging Face models.