Unsupervised Learning for Color Constancy

Abstract

Most digital camera pipelines use color constancy methods to reduce the influence of illumination and camera sensor on the colors of scene objects. The highest accuracy of color correction is obtained with learning-based color constancy methods, but they require a significant amount of calibrated training images with known ground-truth illumination. Such calibration is time consuming, preferably done for each sensor individually, and therefore a major bottleneck in acquiring high color constancy accuracy. Statistics-based methods do not require calibrated training images, but they are less accurate. In this paper an unsupervised learning-based method is proposed that learns its parameter values after approximating the unknown ground-truth illumination of the training images, thus avoiding calibration. In terms of accuracy the proposed method outperforms all statisticsbased and many learning-based methods. An extension of the method is also proposed, which learns the needed parameters from non-calibrated images taken with one sensor and which can then be successfully applied to images taken with another sensor. This effectively enables inter-camera unsupervised learning for color constancy. Additionally, a new high quality color constancy benchmark dataset with 1707 calibrated images is created, used for testing, and made publicly available. The results are presented and discussed. The source code and the dataset are available at

Color Cerberus

“…During the contest we've tried to leverage some publicly available datasets in addition to Cube+ [4]. such as INTEL-TUT [5], NUS [6], Grey Ball [7] and Color Checker [8].…”

The Past and the Present of the Color Checker Dataset Misuse

“…While it may be argued that due to the ill-posedness of the illumination estimation problem this is not a serious issue, it must be stressed that there are no similar issues with other widely used benchmark datasets. As a matter of fact, the case when there are multiple camera sensors involved is treated in the literature as inter-camera color constancy [40], [41]. Therefore, before using the Color Checker dataset to evaluate a method's performance, it should be considered whether it could be affected by learning from images taken from multiple camera sensors whose groundtruth illumination chromaticity distributions are different.…”

Multiple light source dataset for colour research

“…Datasets collected under natural and uncontrolled conditions, for example, REC [9], CubePlus [10], GrayBall [11], NUS [12], are poorly suited for quality evaluation of problems listed above. The estimation of source colour in such datasets is performed using the calibration object (or objects) illuminated by several light sources simultaneously.…”