Tianshu Yu
Papers
1
Total Citations
13
H-Index
1
About
Tianshu Yu is a researcher whose work bridges computer vision and graph-based machine learning, with a particular focus on scene understanding and data association. His key research areas include graph matching, scene parsing, and visual correspondence, where he has developed methods that integrate structured prediction with real-world imagery. His most cited paper, "Scene parsing using graph matching on street-view data" (2016, 13 citations), introduces a novel approach that leverages graph matching to parse complex urban environments from street-level images. This work demonstrates his ability to apply combinatorial optimization to practical vision tasks, enabling more accurate segmentation and recognition of objects in cluttered scenes. Beyond this, Yu's contributions have influenced how researchers model spatial relationships in visual data, offering a structured alternative to purely deep learning-based methods. His impact is evident in the growing adoption of graph-based techniques in scene understanding, and his work continues to inspire students and researchers exploring the intersection of graph theory and computer vision.
Research Focus
Key Achievements
Top Papers
- 1Scene parsing using graph matching on street-view data13 citations · 2016