Stereo Depth from Polarization
Self-supervised stereo depth and surface-normal learning for texture-poor and reflective regions.
As a Research Scientist Intern at Meta Reality Labs, I developed a self-supervised stereo pipeline that uses polarization physics and learned features to improve geometric consistency in texture-poor and reflective regions.
The work jointly predicts depth and surface normals through a multi-objective learning framework. This page intentionally summarizes only the public, résumé-level description of the project.