Max Robotics — Global Service Robot Directory + US Certification

About

Ferdous Sohel is a distinguished researcher whose work spans computer vision, robotics perception, and deep learning, with particular expertise in RGB-D scene understanding, depth estimation, and adversarial machine learning. His most celebrated contribution, "RGB-D Object Recognition and Grasp Detection Using Hierarchical Cascaded Forests" (2017), has garnered over 157 citations and introduced an innovative hierarchical cascaded forest architecture that simultaneously computes object-class and grasp-pose probabilities — a significant advance for robotic manipulation systems. His early work on real-time 6D pose estimation from RGB-D imagery further established his credentials in robotic perception. Sohel has demonstrated a talent for synthesizing complex fields, as evidenced by his comprehensive 2024 survey on deep learning-based monocular depth estimation, which rapidly accumulated 55 citations by consolidating insights from over 500 published works. His research into human interaction prediction and adversarial robustness of 3D object tracking systems reflects a broad and forward-thinking research agenda addressing both fundamental perception challenges and emerging security concerns in autonomous systems. Collectively, his contributions have meaningfully shaped how machines perceive, interpret, and safely interact with the three-dimensional world.