Abstract

Intrinsic image decomposition aims to factorize an image into albedo (reflectance) and shading (illumination) sub-components. Being ill posed and under-constrained, it is a very challenging computer vision problem. There are infinite pairs of reflectance and shading images that can reconstruct the same input. To address the problem, Intrinsic Images in the Wild by Bell et al. provides an optimization framework based on a dense conditional random field (CRF) formulation that considers long-range material relations. We improve upon their model by introducing illumination invariant image descriptors: color ratios. The color ratios and the intrinsic reflectance are both invariant to illumination and thus are highly correlated. Through detailed experiments, we provide ways to inject the color ratios into the dense CRF optimization. Our approach is physics based and learning free and leads to more accurate and robust reflectance decompositions.

© 2021 Optical Society of America

Deep intrinsic decomposition trained on surreal scenes yet with realistic light effects

Hassan A. Sial, Ramon Baldrich, and Maria Vanrell
J. Opt. Soc. Am. A 37(1) 1-15 (2020)

Deep image enhancement for ill light imaging

Rizwan Khan, You Yang, Qiong Liu, Jialie Shen, and Bing Li
J. Opt. Soc. Am. A 38(6) 827-839 (2021)

Experiments on pattern recognition using invariant Fourier–Mellin descriptors

Yunlong Sheng and Henri H. Arsenault
J. Opt. Soc. Am. A 3(6) 771-776 (1986)

Intrinsic decomposition from a single spectral image

Xi Chen, Weixin Zhu, Yang Zhao, Yao Yu, Yu Zhou, Tao Yue, Sidan Du, and Xun Cao
Appl. Opt. 56(20) 5676-5684 (2017)

Autoencoder-based training for multi-illuminant color constancy

Donik Vršnak, Ilija Domislović, Marko Subašić, and Sven Lončarić
J. Opt. Soc. Am. A 39(6) 1076-1084 (2022)