Jaehoon Choi
Papers
1
Total Citations
4
H-Index
1
About
Jaehoon Choi is a leading researcher in computer vision and robotics, specializing in monocular depth estimation and self-supervised learning. His most influential work, "SelfTune: Metrically Scaled Monocular Depth Estimation through Self-Supervised Learning" (2022), addresses a fundamental challenge in the field: recovering metric-scale depth from single images without ground-truth labels. By integrating monocular SLAM with proprioceptive sensors, Choi's algorithm resolves the inherent scale ambiguity that plagues traditional depth prediction methods, enabling robots and autonomous systems to perceive their environment with accurate spatial understanding. This innovation has garnered 4 citations and represents a significant step toward practical, real-world deployment of vision-based navigation. Choi's contributions lie at the intersection of geometric computer vision and deep learning, where he develops algorithms that learn from unlabeled video streams—a paradigm critical for scaling to diverse, unstructured environments. His work is particularly impactful for applications in autonomous driving, augmented reality, and mobile robotics, where metric accuracy is essential for safe interaction. Through SelfTune, Choi has demonstrated that self-supervised learning can achieve the precision previously reserved for supervised methods, opening new avenues for cost-effective, scalable perception systems.
Research Focus
Key Achievements
Top Papers
- 1