Fei Teng

Southwest Jiaotong University, Hunan University, Centre for Artificial Intelligence and Robotics

Papers

Total Citations

H-Index

About

Fei Teng is an emerging researcher at the intersection of computer vision, multimodal perception, and intelligent robotics. His work spans several interconnected domains, including vision-language alignment, multi-object tracking, and RGB-thermal semantic segmentation, with a unifying focus on enhancing the perceptual capabilities of autonomous and robotic systems. Teng's most influential contribution to date is his 2024 work on transformer-based vision-language alignment for robot navigation and question answering, which has garnered 19 citations and reflects growing interest in grounding natural language within spatial visual contexts. This research advances how robotic agents interpret and act upon complex, language-driven instructions in real-world environments. His more recent investigations push perceptual boundaries further. His 2025 work on omnidirectional multi-object tracking addresses a critical limitation in conventional tracking algorithms by leveraging 360° panoramic imagery to capture richer spatial and temporal scene relationships. Complementing this, his exploration of adapting the Segment Anything Model 2 for RGB-thermal segmentation with language guidance demonstrates a commitment to making large foundation models practically deployable in resource-constrained, multi-modal robotic settings. Together, these contributions position Teng as a promising voice in robust, real-world scene understanding research.