Hefeng Wu

Papers

Total Citations

H-Index

About

Hefeng Wu is a computer vision and artificial intelligence researcher whose work spans facial expression recognition, visual tracking, and intelligent human-computer interaction. His research addresses some of the most challenging problems in making machines understand and respond to human behavior in real-world environments. Among his most recognized contributions is his 2021 work on AU-Expression Knowledge Constrained Representation Learning, which tackles a critical limitation of existing deep learning models — their tendency to perform well in controlled laboratory settings but struggle in unconstrained, real-world conditions. By integrating action unit knowledge as a constraint during representation learning, Wu's approach advances the reliability of automatic facial expression recognition, a capability increasingly vital for intelligent robotics and human-machine collaboration. This paper has accumulated 24 citations, reflecting its relevance to the research community. His 2019 work on salient superpixel visual tracking, with 22 citations, demonstrates his breadth across visual understanding tasks, combining graph models with iterative segmentation to improve object tracking robustness. Together, Wu's contributions highlight a commitment to bridging the gap between theoretical deep learning advances and practical, deployable AI systems — making his work particularly valuable for researchers working at the intersection of computer vision, robotics, and affective computing.