ShadowSG: Spherical Gaussian Illumination from Shadows

Hanwei Zhang, Xu Cao, Hiroshi Kawasaki, Takafumi Taketomi

Keywords: illumination from shadow, spherical Gaussian, illumination estimation

TL;DR: We estimate the surrounding illumination of a scene object, parameterized as a mixture of spherical Gaussians, from the appearance of its cast shadows.

Abstract:

This work leverages shadow cues in a scene to infer the surrounding illumination of the shadow-casting object.
Unlike prior works that optimize a discrete environment map, we model scene illumination using a mixture of spherical Gaussians (SGs).
SG illumination provides more intuitive relations to shadow appearance and offers a more compact parameterization compared to discrete environment maps.
To estimate SG parameters, we employ an SG-based, differentiable, closed-form rendering equation to explain the shading of the shadow plane and minimize a photometric loss between the rendered and observed shadow plane shading.
Experiments on synthetic and real-world images under various surrounding illumination demonstrate that our method estimates illumination more accurately than approaches based on discrete environment maps.
With our estimated lighting, consistent shadow effects are realized when blending virtual objects into real-world images.

Submission Number: 271