Color Constancy and Color Consistency Using Dynamic Gamut Adjustment

A lot of color constancy (CC) approaches have been studied to solve color-shifting problems. However, these methods do not consider the color gamut distribution of the image in the color space, leading to color shifting caused by different color temperatures. To address this issue, a novel dynamic gamut adjustment (DGA) method is proposed to reduce the color shifting by mapping the image into a YUV space to obtain the color gamut distribution of the image and filter the UV channel of a single image to select available reference white spots. The
$Y$
channel of these white spots is used as the basis for adjustment. In the experiments, two public datasets (KADID and TID2013) are used to generate the images of six different color temperatures for each image through Adobe Photoshop. The experimental results show that the proposed method can greatly reduce color shifting while maintaining image quality and field distribution under different color temperatures and achieve CC and consistency on two public datasets in comparison with the state-of-the-art approaches.