Presentation

· Contributors · Organizations · Search Program

EyeIR: Single Eye Image Inverse Rendering in the Wild

DescriptionWe propose a method to decompose a single eye image into albedo, shading, specular, normal, and illumination component. We create a diverse synthetic eye dataset and design a synthetic-to-real adaptation framework for self-supervised learning on real images. We also design a method which specifically addresses ambiguities caused by corneal specularity.

Authors

Shijun Liang

Beihang University

Haofei Wang

Peng Cheng Laboratory

Feng Lu

Beihang University

Peng Cheng Laboratory