An image denoising neural network training architecture and a method of training the image denoising neural network

Abstract

An image denoising neural network training architecture includes an image denoising neural network and a clean data neural network, and the image denoising neural network and clean data neural networkshare information between each other.

Description

[0001] Cross References to Related Applications [0002] The present patent application claims priority to U.S. Provisional Patent Application Serial No. 62 / 563,623, filed September 26, 2017, and U.S. Nonprovisional Patent Application Serial No. 15 / 926,301, filed March 20, 2018 of priority, the entire contents of which are incorporated herein by reference. technical field [0003] Aspects of exemplary embodiments of the present disclosure relate to image denoising neural network architectures and methods of training image denoising neural networks. Background technique [0004] When digital images are taken under relatively low lighting conditions, such as at dusk or at night, noise (also known as image noise) is often present in the resulting image. Image noise can appear as random variations in luminance or color information produced by, for example, sensors (e.g., charge-coupled device (CCD) sensors, complementary-metal-oxide semiconductor (CMOS) sensor) and / or the asso...

AI Technical Summary

Problems solved by technology

However, this method does not consider the noise variance smoothing

Images