An automatic image and video colorization algorithm based on pattern continuity

Abstract

This paper proposes a novel method to colorize grayscale images from a reference image. Rather than depending on the comparisons between the grayscale image and the reference image based on the luminance value and neighborhood statistics of each pixel, our method consider pattern continuity and spatial consistency. We handle the pixels in the grayscale image from a higher level and gain the global results, which are better than the results based on the local statistical information. Another advantage of our method is we need less human intervention than the previous methods. This is very beneficial for the video colorization. Our method is focus on colorizing cartoon images and videos. With regard to the cartoon images, using texture analysis based on pattern continuity can gain a better result. For cartoon videos, our method generates scribbles automatically instead of drawing by the user. Once a representative frame has been colorized, all the other frames will be colorized sequentially, which makes a great improvement on automatic colorization.