Papers
1
Total Citations
6
H-Index
1
About
Hyunguk Shon is a researcher advancing efficient computer vision for robotics, with a primary focus on lightweight monocular depth estimation. His most cited work, "Lightweight Monocular Depth Estimation via Token-Sharing Transformer" (2023), addresses a critical bottleneck in mobile robotics: enabling accurate depth perception from a single RGB camera without the computational burden of stereo setups. By introducing a token-sharing mechanism within a transformer architecture, Shon’s design achieves state-of-the-art depth estimation while maintaining a compact, deployable model suitable for resource-constrained platforms. This contribution is vital for autonomous navigation, obstacle avoidance, and scene understanding in drones, service robots, and augmented reality devices. With 6 citations already, the paper is gaining traction as a practical solution for real-world deployment. Shon’s work bridges the gap between high-performance deep learning and the stringent efficiency demands of embedded systems, positioning him as a key figure in the push toward accessible, real-time 3D perception. His research continues to inspire new directions in token-efficient transformers and lightweight vision models for robotics.
Research Focus
Key Achievements
Top Papers
- 1Lightweight Monocular Depth Estimation via Token-Sharing Transformer6 citations · 2023