Xingwang Ma
Papers
1
Total Citations
2
H-Index
1
About
Xingwang Ma is a rising researcher in the field of computer vision and robotics, with a focused expertise in dense Simultaneous Localization and Mapping (SLAM) and 3D scene reconstruction. His work centers on enhancing the performance and efficiency of real-time mapping systems, particularly for applications in augmented reality and autonomous navigation. Ma’s most notable contribution, detailed in his 2024 paper “Enhancing SplaTAM Performance Through Dynamic Learning Rate Decay and Optimized Keyframe Selection,” introduces critical optimizations to the SplaTAM framework, which leverages 3D Gaussian representations for high-quality reconstruction from RGB-D cameras. By implementing a dynamic learning rate decay strategy and a refined keyframe selection process, his research significantly improves the accuracy and stability of online tracking and mapping. Although early in his career, with his flagship paper already garnering 2 citations, Ma’s work addresses a fundamental bottleneck in dense SLAM—balancing computational efficiency with reconstruction fidelity. His innovative approach to optimizing neural rendering pipelines positions him as a promising contributor to the next generation of spatial AI systems.
Research Focus
Key Achievements
Top Papers
- 1