Anuja Dawane

Papers

Total Citations

H-Index

About

Anuja Dawane is a computer vision researcher whose work focuses on advancing human behavior understanding from visual data, with particular emphasis on body orientation estimation and 3D pose analysis. Her most impactful contribution is the development of MEBOW (Monocular Estimation of Body Orientation in the Wild), a method that robustly estimates body orientation from single images even under challenging conditions such as low resolution, occlusion, or ambiguous body configurations. This work is critical for applications in robotics, autonomous driving, and human-computer interaction, where understanding a person's facing direction provides essential behavioral cues when full 3D pose estimation is infeasible. To support this research, Dawane introduced the COCO-MEBOW dataset, which has become a valuable benchmark for the field. Her paper on this work has garnered over 40 citations, reflecting its significance in enabling practical, real-world deployment of orientation-aware systems. By tackling the difficult problem of orientation estimation in uncontrolled environments, Dawane has made a notable contribution to making computer vision systems more perceptive and context-aware in dynamic, unconstrained settings.