Changhai Man

Southern University of Science and Technology

Papers

Total Citations

H-Index

About

Changhai Man is a researcher at the forefront of efficient deep learning for robotic perception and autonomous systems. His primary research areas center on visual simultaneous localization and mapping (VSLAM), visual odometry, and loop closure detection, with a specific focus on deploying these computationally intensive algorithms on edge devices. Man’s major contributions lie in developing novel, compressed neural network architectures that maintain high performance while drastically reducing model size and computational cost. His work on ATFVO introduced an attentive tensor-compressed LSTM model that leverages optical flow features for monocular visual odometry, enabling real-time deployment on resource-constrained hardware. Building on this, his TT-LCD framework pioneered a tensorized-Transformer approach for loop closure detection in robotic VSLAM, directly addressing the critical challenge of correcting drift and accumulated errors in long-term navigation. These innovations are vital for advancing autonomous driving, intelligent robotics, and metaverse applications. With his papers accumulating citations and his focus on bridging the gap between state-of-the-art deep learning and practical edge deployment, Changhai Man is establishing himself as a key contributor to making robust, real-time visual perception a reality for next-generation autonomous systems.