Md Jahidul Islam
Papers
8
Total Citations
143
H-Index
6
About
Md Jahidul Islam is a pioneering researcher at the intersection of underwater robotics, computer vision, and deep learning, with a focused mission to enable intelligent visual perception in challenging aquatic environments. His work addresses some of the most demanding problems in marine robotics, including monocular depth estimation, salient object detection, image enhancement, and autonomous navigation in visually degraded settings. Islam's most impactful contributions include UDepth (47 citations), a fast end-to-end deep learning pipeline for monocular depth estimation tailored to low-cost underwater robots, and SVAM-Net (41 citations), a sophisticated saliency-guided visual attention model enabling autonomous underwater robots to identify and prioritize objects of interest. His earlier work on mixed-domain biological motion tracking (23 citations) demonstrated innovative human-robot interaction techniques for diver-following robots, blending spatial and frequency-domain analysis. His Deep SESR framework tackled simultaneous image enhancement and super-resolution for real-time underwater vision applications. More recently, Islam has pushed into autonomous cave exploration, developing weakly supervised caveline detection systems for AUVs navigating hazardous underwater caves. Collectively, his research — spanning fundamental perception algorithms to edge-deployable systems — is shaping the future of intelligent underwater robotics and ocean exploration.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Computer Vision Applications in Underwater Robotics and Oceanography8 citations · 2024
- 6
- 7
- 8