Neural Networks-based Image Denoising Methods
Mengke Wang · Preprints.org · 2023
Image denoising has been one of the important problems in the field of computer vision, and it has a wide range of practical value in many applications, such as medical image processing, image enhancement, and computational photography. Traditional image denoising methods are usually based on hand-designed features and filters, but these methods perform poorly under complex noise and image structures. In recent years, the rapid development of neural network technology has revolutionized the image-denoising task. This paper introduces the knowledge about neural networks and image denoising, explores the impact of neural networks on image denoising, and how is it possible to denoise images by neural networks. It also summarises other image-denoising methods and finally points out the challenges and problems faced by image-denoising at present. Some possible new development directions are proposed to provide new solutions for image-denoising researchers and to promote the development of the field.