Alisha Sharma

United States Naval Research Laboratory, Naval Research Laboratory Laboratories for Computational Physics & Fluid Dynamics

Papers

Total Citations

H-Index

About

Alisha Sharma is a researcher advancing the frontier of computer vision, with a focus on unsupervised learning for panoramic depth estimation and ego-motion. Her work addresses a critical challenge: enabling machines to perceive 3D structure from 360° video without requiring expensive labeled data. Sharma’s key contribution is a convolutional neural network model that learns depth and camera motion directly from cylindrical panoramic footage. This innovation has direct applications in virtual reality, 3D modeling, and autonomous robotic navigation, where understanding full-surround environments is essential. Her most-cited paper (6 citations) and accompanying video demonstration showcase how her method outperforms prior approaches on street-level 360° data. By removing the need for ground-truth depth, Sharma’s unsupervised framework makes panoramic perception more scalable and practical. Her work is foundational for immersive VR experiences and self-driving systems that rely on omnidirectional sensors. Sharma’s research sits at the intersection of deep learning, geometry, and real-world deployment, offering a promising path toward robust, label-free spatial understanding.