Papers
9
Total Citations
263
H-Index
6
About
Sifan Zhou is a robotics and computer vision researcher whose work centers on 3D object tracking, point cloud processing, autonomous driving perception, and efficient deep learning deployment. Zhou has made particularly influential contributions to LiDAR-based 3D single object tracking, pioneering methods that push the boundaries of real-time performance and robustness. His 2020 work, *3D-SiamRPN*, introduced an end-to-end learning framework for tracking objects directly from raw point clouds, earning 82 citations, while his landmark *Point-Track-Transformer (PTT)* module (2021, 83 citations) demonstrated how transformer architectures—leveraging self-attention and position encoding—could dramatically improve feature extraction for 3D tracking. A follow-up real-time transformer-based tracker (2022) extended this line of inquiry to address sparse, occluded point clouds at long range, accumulating 59 citations. More recently, Zhou has broadened his research scope to include model efficiency, developing post-training quantization strategies for LiDAR detectors (*LiDAR-PTQ*, 2024) and quantization-aware pillar feature encoders (*PillarHist*, 2025) suited for edge deployment. His exploration of diffusion-based stereo matching and foundation models for IMU odometry signals an exciting expansion toward generalizable, cross-modal robotic perception systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Real-Time 3D Single Object Tracking With Transformer59 citations · 2022
- 4Point Siamese Network for Person Tracking Using 3D Point Clouds15 citations · 2019
- 5
- 6
- 7
- 8
- 9