A Review of Gray-scale Image Recoloring Methods With Neural Network Based Model

Gray-scale image recolor or computer-assisted color restoration is a computer-assisted process of adding color in-formation to grayscale photographs, which is crucial in perceiving visual information by humans. An important feature of this technique is the ability to restore the original style of old images. Neural network-based black and grayscale image coloring technology is gradually becoming a prevalent topic in current research with the development of computer vision and data science. Models with DNN, CNN, GAN, and DCGAN were successively proposed in the recent few years, indicating a fresh-new hot spot of image processing. Several popular image re-coloring methods have been discussed in this essay, including a review of recent research and then a detailed analysis of the current deployment method. Finally, the essay compares the existing application, demonstrating several real-world examples and providing information and statistics on how to apply this technique to the industry. We also summarized current limitations and future trends of this research field to provide an outlook for other researchers.