Geonmo Yang
Papers
4
Total Citations
34
H-Index
3
About
Geonmo Yang is a researcher specializing in underwater and maritime visual perception, with a particular focus on image enhancement, depth estimation, and robust sensing for robotic systems operating in challenging environments. His work addresses fundamental problems in degraded visual conditions — including scattering, absorption, and haze — that severely limit the effectiveness of autonomous systems in aquatic settings. Yang's most notable contribution, "Joint-ID" (2023, 19 citations), introduced a transformer-based framework for simultaneously enhancing underwater images and estimating depth, demonstrating the power of joint learning in overcoming the compounded degradations unique to underwater imaging. Building on this foundation, his 2024 work TRIDENT extended this philosophy to a triple-task learning system integrating dehazing, depth estimation, and uncertainty quantification, advancing the reliability of underwater robotic perception. His development of the PoLaRIS dataset (2025, 9 citations) further reflects a commitment to community-wide progress, providing a dedicated benchmark for maritime object detection and tracking in real-world canal environments. Additional work leveraging Gaussian Processes for sparse depth-guided dehazing highlights his versatility across probabilistic and deep learning methodologies. Collectively, Yang's research meaningfully advances the safety and autonomy of marine robotic navigation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4