Shaocheng Jia
Papers
4
Total Citations
52
H-Index
3
About
Shaocheng Jia is a computer vision researcher whose work centers on self-supervised depth estimation, 3D scene reconstruction, and ego-motion estimation, with particular emphasis on applications in autonomous driving, robotics, and intelligent transportation systems. Jia has made notable contributions to the challenge of extracting accurate spatial information from monocular camera systems — a problem considered inherently ill-posed due to the loss of depth cues in 2D imagery. His research has progressively advanced the field by addressing key limitations such as global context extraction and geometric consistency, culminating in frameworks that leverage transformer-based global perception alongside geometric smoothness constraints to improve depth prediction without requiring labeled training data. Among his most recognized contributions, his 2022 paper on self-supervised depth estimation leveraging global perception has accumulated 21 citations, while his 2021 work on self-supervised 3D reconstruction and ego-motion estimation via monocular video has garnered 20 citations — together representing a coherent research trajectory pushing the boundaries of label-free scene understanding. His earlier hybrid neural network approach further demonstrated creative integration of architectural innovations to tackle depth ambiguity. Collectively, Jia's research offers meaningful advances for real-world autonomous systems operating under practical data constraints.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4