About Me

I am currently a first-year master's student at Central South University, under the supervision of Prof. Shichao Kan. Before that, I received my bachelor's degree from Wuhan Institute of Technology in 2026, where I worked closely with Prof. Lu Zou.

My recent research focuses on Multimodal Large Language Models (MLLMs), Video Understanding, and Spatial Intelligence. If you are interested in collaborating with me, please feel free to contact me via xiaozhang0479@gmail.com or WeChat!

News

• 🎉 2026.07: REF-Pose is accepted by ACM MM 2026.
• 🎉 2026.04: I join SenseTime Research as a research intern.
• 🎉 2025.04: My first work, INKL-Pose, is available on arXiv.

Publications

Rethinking Visual Token Compression for Video Large Language Models: A Simple Yet Strong Baseline

Xiao Zhang, Wang Zeng, Sheng Jin, Wentao Liu, Chen Qian, Shichao Kan

arXiv Preprint 2026

[Paper] [Code]

Thinking with Cameras: Active Visual Reasoning via Dynamic Viewpoint Control for Surveillance Video Understanding

Xiao Zhang, Wang Zeng, Sheng Jin, Wentao Liu, Chen Qian, Shichao Kan

arXiv Preprint 2026

[Paper] [Code] [Model & Datasets]

Rotation-Equivariant and Spatially Factorized Keypoint Modeling for Category-Level Object Pose Estimation

Lu Zou, Xiao Zhang, Zhangjin Huang, Guoping Wang

Student First Author

The 34th ACM International Conference on Multimedia (ACM MM 2026)

[Paper]

Instance-Adaptive Keypoint Learning with Local-to-Global Geometric Aggregation for Category-Level Object Pose Estimation

Xiao Zhang, Lu Zou, Tao Lu, Yuan Yao, Zhangjin Huang, Guoping Wang

arXiv Preprint 2025

[Paper] [Code]

Education

• 2026.09 - 2029.06 | Central South University Master's Degree in Computer Technology.
• 2022.09 - 2026.06 | Wuhan Institute of Technology Bachelor's Degree in Artificial Intelligence. Rank: 2/106.