A New Debleeding Algorithm for Image colorization

Abstract

Colorization is the process of adding color to grayscale images using computer algorithms. There are several approaches to color an image, including precise image segmentation algorithms, deep learning algorithms, and manually colored local color expansion methods. However, one common problem with these approaches is the occurrence of “color bleeding”, where colors from one region of the image spill over into adjacent regions. In this paper, we propose a new manually colored local color expansion algorithm that considers the intensity value difference and distance difference between the central pixel in the window and its neighbor pixels comprehensively. Combined with side window filtering, our algorithm significantly reduces the occurrence of colors bleeding at the edges of the colored image. The experiments demonstrate the effectiveness of the proposed algorithm.