Shizhong Han

Qualcomm (United Kingdom)

Papers

2

Total Citations

56

H-Index

2

About

Shizhong Han is a researcher at the forefront of 3D computer vision and multimodal learning, with a particular focus on bridging the gap between natural language understanding and 3D scene perception. His most recognized contribution, **PartSLIP**, addresses one of the field's persistent challenges: enabling accurate part-level segmentation of 3D point clouds without requiring extensive labeled training data. By leveraging pretrained image-language models, Han's approach offers a compelling low-shot alternative to conventional supervised pipelines, significantly reducing the costly burden of fine-grained 3D annotation — a bottleneck that has long hindered progress in robotics and embodied AI applications. PartSLIP has garnered over 50 citations since its 2023 publication, reflecting strong community interest in scalable, generalizable solutions for 3D understanding. The work demonstrates Han's ability to creatively repurpose large-scale vision-language models for structured geometric tasks, a direction that is increasingly influential as foundation models expand into spatial reasoning. For students and researchers working at the intersection of 3D perception, robotics, and multimodal AI, Han's research represents a practical and innovative pathway toward building systems that understand the physical world with minimal supervision.

Research Focus

Key Achievements

2
H-Index
2
Papers
56
Total Citations
28
Avg Citations/Paper
🏆 Most Cited Paper
PartSLIP: Low-Shot Part Segmentation for 3D Point Clouds via Pretrained Image-Language Models
53 citations · 2023
📈 Most Prolific Year: 2023 (1 Papers)
🤝 Key Collaborators: 8
🏛 Institutions: Qualcomm (United Kingdom)

Top Papers

  1. 1
  2. 2

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago