📝 Publications
Machine Vision
Robust Point Cloud Registration Framework Based on Deep Graph Matching
Kexue Fu, Shaolei Liu, Xiaoyuan Luo, Manning Wang
- Propose a novel deep graph matching based framework for point cloud registration.
- Propose a module based on deep graph matching to calculate graph edge.
PointMBF: A Multi-scale Bidirectional Fusion Network for Unsupervised RGB-D Point Cloud Registration
Mingzhi Yuan, Kexue Fu, Zhihao Li, Yucong Meng, Manning Wang
- Propose a network implementing multi-scale bidirectional fusion strategy.
- Our method achieves new state-of-the-art performance on ScanNet and 3DMatch.
PI-NeRF: A Partial-Invertible Neural Radiance Fields for Pose Estimation
Zhihao Li, Kexue Fu, Haoran Wang, Manning Wang
- Pose Estimation Without Initialization and Iterative Optimization.
- Significant Speed Improvement with Competitive Accuracy.
Zhiwei Yang, Kexue Fu, Minghong Duan, Linhao Qu, Shuo Wang, Zhijian Song
- Novel ‘Separate and Conquer’ Scheme for Co-occurrence Problem.
- Enhanced Semantic Representation with Multi-Granularity Knowledge Contrast.
🏥 Medical LMs
FAST: A Dual-tier Few-Shot Learning Paradigm for Whole Slide Image Classification
Kexue Fu, Xiaoyuan Luo, Linhao Qu, Shuo Wang, Ying Xiong, Ilias Maglogiannis, Longxiang Gaom, Manning Wang
- A novel and efficient dual-tier few-shot learning paradigm for WSI classification.
- Approache the accuracy of fully supervised methods with only 0.22% annotation costs.
Linhao Qu, Xiaoyuan Luo, Kexue Fu, Manning Wang, Zhijian Song
- Few-Shot Weakly Supervised Learning for WSI Classification
- Two-Level Prompt Learning with Vision-Language Models.
💊 AI4Science
ProteinMAE: Masked Autoencoder for Protein Surface Self-supervised Learning
Mingzhi Yuan, Ao Shen, Kexue Fu, Jiaming Guan, Yingfan Ma, Qin Qiao, Manning Wang
- A Self-Supervised Framework for Protein Surface Representation.
- Efficiency and Competitive Performance.