Cha Zhang

Microsoft (Germany), Zhejiang University, Microsoft (United States)

Papers

Total Citations

H-Index

About

Cha Zhang is a researcher whose work spans computer vision, model compression, and multimedia technologies. He is perhaps best known for his contributions to neural network efficiency, particularly through his development of LeGR (Learned Global Ranking), a novel approach to filter pruning in convolutional neural networks. Unlike prior methods that require users to manually specify target model complexity, Zhang's LeGR framework introduces a learned global ranking strategy that automates and optimizes the pruning process — a significant step forward in making deep learning models more practical for real-world deployment. This work, published across 2019 and 2020, has garnered notable attention within the model compression community, accumulating over 40 combined citations. Beyond neural network efficiency, Zhang has also made contributions to robotics and sensing, including a robust optical/inertial data fusion system for robot manipulator motion tracking that improves upon traditional optical tracking systems through IMU integration. His editorial leadership in 3D imaging techniques further reflects a broad expertise in multimedia applications, encompassing sensing, transmission, and visualization technologies. Together, his body of work demonstrates a versatile research profile bridging efficient deep learning, computer vision, and applied sensing systems.