H. Mohaghegh

Isfahan University of Technology, The University of Western Australia

Papers

Total Citations

H-Index

About

H. Mohaghegh is a computer vision researcher whose work focuses on solving fundamental problems in 3D scene understanding and face reconstruction from single images. Their primary research areas include monocular depth estimation, 3D face reconstruction, and reinforcement learning for computer vision tasks. Mohaghegh’s most significant contribution is the development of a novel stacked generalization model that aggregates rich depth-aware features for single image depth estimation, a paper that has garnered 23 citations and addresses the ill-posed problem of inferring 3D structure from 2D inputs—a critical capability for applications in robotics, autonomous navigation, and 2D-to-3D conversion. Their earlier work on joint local-global features for depth estimation laid the groundwork for this approach, while their most recent research introduces a label-efficient reinforcement learning framework for 3D face reconstruction, designed to improve robustness against occlusions and noise in real-world human-robot interaction systems. This innovative method reduces the need for extensive labeled training data, making it particularly valuable for applications in automatic face authentication and entertainment. Mohaghegh’s work consistently pushes toward more practical, data-efficient solutions for 3D vision challenges.