Generative Adversarial Networks Used in Color Modulation
Conference: ICMLCA 2021 - 2nd International Conference on Machine Learning and Computer Application
12/17/2021 - 12/19/2021 at Shenyang, China
Proceedings: ICMLCA 2021
Pages: 9Language: englishTyp: PDF
Personal VDE Members are entitled to a 10% discount on this title
Authors:
Huang, Ruirong (Shanghai Jiao Tong University University of Michigan Joint Institute, Shanghai Jiao Tong University, Shanghai, China)
Abstract:
Nowadays, color modulation has become more and more important in the movie industry. Color modulation agents were asked to adjust the color style of every different image shot by photographers in various situations and therefore composed of various color styles, into a coherent style throughout the whole movie, in order to make the audience feel smooth and consistent when watching the movie. However, present techniques for color modulation were still based on the idea to build a mapping function from a specific image to another (or in microcosmic view, from pixels to pixels), which means that if facing lots of raw images with various color styles, the agent would have to do color modulation for all these different pictures one by one, and this leads to a great waste of time and even enthusiasm. Therefore, a new color modulation method using GAN (Generative Adversarial Network) is presented. By learning from the images of target color style, it could generate a model that would map from any original color style into the target style, which will save much labor cost.