Weixiang Liu
Papers
2
Total Citations
5
H-Index
1
About
Weixiang Liu is a researcher advancing the frontiers of robot perception and continual learning. His work centers on robust keypoint detection for pose estimation and multi-domain pattern analysis under federated constraints. In his highly cited paper "G-SAM: A Robust One-Shot Keypoint Detection Framework for PnP Based Robot Pose Estimation" (2023, 4 citations), Liu introduced a novel one-shot framework that significantly improves the accuracy and robustness of perspective-n-point (PnP) based pose estimation—a critical capability for autonomous manipulation and augmented reality. More recently, his 2025 work "Family-based continual learning for multi-domain pattern analysis in federated frameworks with GCN and ViT" (1 citation) pioneers a family-based continual learning strategy that integrates Graph Convolutional Networks (GCN) and Vision Transformers (ViT) to enable adaptive, privacy-preserving pattern recognition across evolving domains. This contribution addresses the pressing challenge of catastrophic forgetting in federated learning systems. Liu’s research is notable for its practical impact on real-world robotics and its theoretical elegance in merging geometric reasoning with modern deep learning architectures. His work continues to inspire new directions in one-shot learning and lifelong machine learning.
Research Focus
Key Achievements
Top Papers
- 1
- 2