Papers
2
Total Citations
113
H-Index
2
About
Dr. Fangyu Hu is a leading researcher in 3D computer vision and point cloud processing, with a focus on developing robust, efficient representations for low-cost sensor data. Their major contributions center on creating novel binary feature descriptors that enable reliable registration and shape analysis from noisy, incomplete 3D scans. Hu’s most influential work, the "Local Voxelized Structure" (LVS) descriptor (2018, 100 citations), introduced a groundbreaking approach to encoding local geometry into compact binary strings, dramatically improving the speed and accuracy of point cloud alignment for consumer-grade sensors like Kinect and LiDAR. This work has become a foundational reference for real-time 3D mapping and robotics applications. Hu further advanced the field with the "Multi-View Silhouette" (MVS) feature (2018, 13 citations), which leverages silhouette information from multiple perspectives to create distinctive shape representations. Their research bridges the gap between theoretical geometry and practical sensor limitations, enabling robust performance in challenging environments with noise, occlusion, and sparse data. Hu’s contributions are widely cited in autonomous navigation, cultural heritage preservation, and augmented reality, making them a key figure in democratizing 3D perception for real-world systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2