Xinhua Xu

Sun Yat-sen University, Peking University

Papers

Total Citations

H-Index

About

Xinhua Xu is an emerging researcher specializing in computer vision and robotic perception, with a particular focus on semantic scene understanding for autonomous systems. Their work centers on developing intelligent segmentation frameworks that enable robots to interpret complex indoor environments with greater accuracy and efficiency. Xu's most notable contribution, "Interactive Efficient Multi-Task Network for RGB-D Semantic Segmentation" (2023, 7 citations), addresses a critical limitation in robotic perception by integrating depth information alongside traditional RGB data. By designing a multi-task architecture that balances performance gains with computational efficiency, Xu tackled the real-world challenge of deploying advanced segmentation models in time-sensitive robotic applications. Building on this foundation, their more recent work on multi-robot collaborative segmentation (2025) pushes the boundaries further by exploring how multiple robotic agents can collectively overcome the perceptual limitations inherent in single-robot systems through multiplex interactive learning. This contribution is particularly forward-looking, recognizing that the future of robust indoor scene understanding may lie in cooperative intelligence rather than isolated processing. Though early in their research career, Xu demonstrates a clear and consistent vision: bridging the gap between deep learning-based perception and practical robotic deployment in real-world environments.