Zishu Liu
Papers
1
Total Citations
6
H-Index
1
About
Zishu Liu has made significant contributions to 3D computer vision, with a primary focus on point cloud processing for object classification. His most cited work, "Pointwise CNN for 3D Object Classification on Point Cloud" (2021), addresses a fundamental challenge in the field: applying convolutional neural networks to the irregular, unordered structure of raw point cloud data. Liu’s approach introduces a pointwise convolutional architecture that enables direct processing of 3D point clouds without requiring voxelization or multi-view projections, making it more efficient for applications in 3D modeling, face recognition, and robotic perception. This work has garnered 6 citations, reflecting its relevance to researchers tackling similar problems in 3D deep learning. Liu’s research sits at the intersection of geometric deep learning and practical computer vision, aiming to bridge the gap between traditional CNN frameworks and the unique demands of 3D data. His contributions are particularly valuable for students and researchers exploring efficient, end-to-end methods for real-time 3D object recognition in autonomous systems and robotics.
Research Focus
Key Achievements
Top Papers
- 1Pointwise CNN for 3D Object Classification on Point Cloud6 citations · 2021