Color correction for mobile device camera image using a modified image formation model

Abstract

In an image capture, a scene with non-uniform illumination has an influence on the image quality, especially the contrast and details in dark regions. For this reason, this paper presents a color correction method using a modified image formation model in which an image is divided into three components: global illumination, local illumination, and reflectance. The global illumination is obtained by using Gaussian filtering of the original image, and then the local illumination is estimated through the adaptive smoothing method. The reflectance is estimated by dividing the original image by the estimated global illumination and local illumination to remove the dominant chromaticity. After performing the power function to compress the dynamic range, the output image is obtained from their product based on sRGB color representation. The experimental results show that the proposed method yield better performance of color correction over the conventional methods.