Bikramjot Hanzra

Papers

Total Citations

H-Index

About

Bikramjot Hanzra is a computer vision researcher whose work focuses on bridging the gap between 2D visual data and 3D spatial understanding, particularly in challenging real-world conditions. His most cited paper, "MEBOW: Monocular Estimation of Body Orientation in the Wild" (2020, 40 citations), addresses a critical problem in autonomous driving and robotics: inferring body orientation from a single image when traditional 3D pose estimation fails due to low resolution, occlusion, or ambiguous body parts. Hanzra introduced the COCO-MEBOW dataset and framework, which enables robust orientation estimation directly from monocular images, offering a practical solution for scenarios where full 3D reconstruction is infeasible. This work has been influential in advancing human-centric perception systems, providing essential visual cues for safe navigation and human-robot interaction. By tackling the limitations of existing methods in uncontrolled environments, Hanzra's contributions help make autonomous systems more reliable in real-world settings. His research continues to impact the fields of embodied AI and intelligent transportation, demonstrating how targeted geometric reasoning can overcome the constraints of imperfect visual data.