Research Progress in Image Denoising Algorithms Based on Deep Learning

Abstract

Images constitute the main source of information to people. Digital images are vital means of obtaining, processing, analyzing, and sharing information in the era of information. Now they have been deeply incorporated into every aspect of people’s production and life, generating considerable social and economic benefits. Thus improving the quality of images and reducing the negative impact of image noises to subsequent image processing have been two important research topics. Image processing technology has been combined with such research fields as cognitive psychology, machine learning, machine vision, and deep learning in recent years, which lead to an unprecedented development level and breakthroughs. Therefore, the study of image denoising technology has profound theoretical significance and promising prospects in practical application. The paper mainly discusses the development course of applying deep learning technology to the image denoising field. Meanwhile, it introduces various classical denoising algorithms and focuses on the thoughts, advantages, and disadvantages of each algorithm. Furthermore, it discusses the challenges faced by deep learning in the denoising field and puts forward possible solutions.