Skip to main content

Research Repository

Advanced Search

Outputs (93)

Stable Hand Pose Estimation under Tremor via Graph Neural Network
Presentation / Conference Contribution
Leng, Z., Chen, J., Shum, H. P., Li, F. W., & Liang, X. (2021, March). Stable Hand Pose Estimation under Tremor via Graph Neural Network. Presented at 2021 IEEE Virtual Reality and 3D User Interfaces (VR), Lisboa

Hand pose estimation, which predicts the spatial location of hand joints, is a fundamental task in VR/AR applications. Although existing methods can recover hand pose competently, the tremor issue occurring in hand motion has not been completely solv... Read More about Stable Hand Pose Estimation under Tremor via Graph Neural Network.

Two-Person Interaction Augmentation with Skeleton Priors
Presentation / Conference Contribution
Li, B., Ho, E. S. L., Shum, H. P. H., & Wang, H. (2024, June). Two-Person Interaction Augmentation with Skeleton Priors. Presented at 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, Washington

Close and continuous interaction with rich contacts is a crucial aspect of human activities (e.g. hugging, dancing) and of interest in many domains like activity recognition, motion prediction, character animation, etc. However, acquiring such skelet... Read More about Two-Person Interaction Augmentation with Skeleton Priors.

From Category to Scenery: An End-to-End Framework for Multi-Person Human-Object Interaction Recognition in Videos
Presentation / Conference Contribution
Qiao, T., Li, R., Li, F. W. B., & Shum, H. P. H. (2024, December). From Category to Scenery: An End-to-End Framework for Multi-Person Human-Object Interaction Recognition in Videos. Presented at Proceedings of the 2024 International Conference on Pattern Recognition, Kolkata, India, 2024., Kolkata, India

Video-based Human-Object Interaction (HOI) recognition explores the intricate dynamics between humans and objects, which are essential for a comprehensive understanding of human behavior and intentions. While previous work has made significant stride... Read More about From Category to Scenery: An End-to-End Framework for Multi-Person Human-Object Interaction Recognition in Videos.

Depth-Aware Endoscopic Video Inpainting
Presentation / Conference Contribution
Xiatian Zhang, F., Chen, S., Xie, X., & Shum, H. P. (2024, October). Depth-Aware Endoscopic Video Inpainting. Presented at 27th International Conference on Medical Image Computing and Computer Assisted Intervention, Marrakesh, Morocco

MxT: Mamba x Transformer for Image Inpainting
Presentation / Conference Contribution
Chen, S., Atapour-Abarghouei, A., Zhang, H., & Shum, H. P. H. (2024, November). MxT: Mamba x Transformer for Image Inpainting. Presented at BMVC 2024: The 35th British Machine Vision Conference, Glasgow, UK

MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment
Presentation / Conference Contribution
Zhou, K., Wang, L., Zhang, X., Shum, H. P. H., Li, F. W. B., Li, J., & Liang, X. (2024, September). MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment. Presented at ECCV 2024: The 18th European Conference on Computer Vision, Milan, Italy

Action Quality Assessment (AQA) evaluates diverse skills but models struggle with non-stationary data. We propose Continual AQA (CAQA) to refine models using sparse new data. Feature replay preserves memory without storing raw inputs. However, the mi... Read More about MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment.

Neural-code PIFu: High-fidelity Single Image 3D Human Reconstruction via Neural Code Integration
Presentation / Conference Contribution
Liu, R., Remagnino, P., & Shum, H. P. (2024, December). Neural-code PIFu: High-fidelity Single Image 3D Human Reconstruction via Neural Code Integration. Presented at 2024 International Conference on Pattern Recognition, Kolkata, India

We introduce neural-code PIFu, a novel implicit function for 3D human reconstruction, leveraging neural codebooks, our approach learns recurrent patterns in the feature space and reuses them to improve current features. Many existing methods predict... Read More about Neural-code PIFu: High-fidelity Single Image 3D Human Reconstruction via Neural Code Integration.