Decolorization algorithm based on contrast pyramid transform fusion

Nana Yu, Jinjiang Li, Zhen Hua

Abstract: The purpose of decolorization is to convert a color image into a grayscale image while maintaining the contrast and structural characteristics of the original color image as much as possible. This paper proposes a decolorization method based on contrast pyramid transform fusion. Image fusion based on contrast tower decomposition is performed on different spatial frequency bands, so it is possible to obtain a fusion effect closer to human visual characteristics. First, we extract the R, G and B channel images of the original color image and the initial grayscale image. Then, we perform Laplacian pyramid fusion of the R, G and B channel images of the original color image with the initial grayscale image to obtain three images. Finally, the three images obtained are reconstructed by contrast pyramid fusion, and our result image is obtained. The visual comparison and quantitative experimental results show that the proposed contrast pyramid fusion decolorization method is more robust, can fully maintain image details and other information, and reduce the distortion in the process of fusion and decolorization.