Jiaming Sun

Zhejiang University, Beijing Institute of Technology

Papers

Total Citations

H-Index

About

Jiaming Sun is a leading researcher in computer vision and robotics, with a focus on depth estimation and visual odometry. His most impactful work introduces a groundbreaking paradigm for metric depth estimation through "Prompt Depth Anything," which, for the first time, incorporates prompting into depth foundation models. This innovation enables accurate, 4K-resolution metric depth estimation using low-cost inputs, achieving 15 citations since 2025 and opening new possibilities for autonomous systems and augmented reality. In earlier work, Sun addressed the fundamental scale-ambiguity challenge in monocular visual odometry by developing a fully scaled monocular direct sparse odometry system with a distance constraint. This approach leverages external distance measurements—such as from GPS or wheel encoders—to restore scale without relying on expensive inertial measurement units or multi-camera setups. Sun’s contributions are pivotal for advancing cost-effective, high-precision 3D perception, with applications spanning robotics, autonomous driving, and spatial computing. His research continues to push the boundaries of what is achievable with monocular vision, making him a key figure in the field.