Conditional GANs for Multi-Illuminant Color Constancy: Revolution or Yet Another Approach?
This work addresses color constancy and shadow removal for computer vision applications, but it appears incremental as it applies an existing GAN framework to a new domain.
The authors tackled multi-illuminant color constancy and shadow removal by proposing a generative end-to-end algorithm based on conditional GANs, and they created a dataset of approximately 6,000 synthetic image pairs to address the lack of training data.
Non-uniform and multi-illuminant color constancy are important tasks, the solution of which will allow to discard information about lighting conditions in the image. Non-uniform illumination and shadows distort colors of real-world objects and mostly do not contain valuable information. Thus, many computer vision and image processing techniques would benefit from automatic discarding of this information at the pre-processing step. In this work we propose novel view on this classical problem via generative end-to-end algorithm based on image conditioned Generative Adversarial Network. We also demonstrate the potential of the given approach for joint shadow detection and removal. Forced by the lack of training data, we render the largest existing shadow removal dataset and make it publicly available. It consists of approximately 6,000 pairs of wide field of view synthetic images with and without shadows.