CVDec 1, 2017

Unsupervised Learning for Color Constancy

arXiv:1712.00436v4112 citations
AI Analysis

This addresses a major bottleneck in digital camera pipelines by reducing calibration time for color constancy, benefiting camera manufacturers and users.

The paper tackles the need for calibrated training images in learning-based color constancy by proposing an unsupervised method that approximates unknown ground-truth illumination, outperforming statistics-based and many learning-based methods in accuracy. It also introduces a new benchmark dataset with 1707 calibrated images and enables inter-camera learning.

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 statistics-based 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 http://www.fer.unizg.hr/ipg/resources/color_constancy/.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes