Abstract

A method is investigated for training neural networks for image enhancement, based on using information from the features of neural networks trained in image classification. Experiments are performed to identify the optimal loss function that achieves maximum precision in the classification of images superposed with noise or blurring. The dependence of the best configuration of training parameters on the type of detrimental influence and target is demonstrated. This is the first study, to our knowledge, to compare the influence of such a loss function on the precision of the restoration and recognition with the utilization of a single classifier trained under the influence of distorting factors. We show that it is reasonable to correct some simple distortions “outside” the classifier, while others are better corrected “inside.”

© 2021 Optical Society of America

Physics-informed deep neural network for image denoising

Emmanouil Xypakis, Valeria de Turris, Fabrizio Gala, Giancarlo Ruocco, and Marco Leonetti
Opt. Express 31(26) 43838-43849 (2023)

Deblurring adaptive optics retinal images using deep convolutional neural networks

Xiao Fei, Junlei Zhao, Haoxin Zhao, Dai Yun, and Yudong Zhang
Biomed. Opt. Express 8(12) 5675-5687 (2017)

Physics-informed neural network for polarimetric underwater imaging

Haofeng Hu, Yilin Han, Xiaobo Li, Liubing Jiang, Li Che, Tiegen Liu, and Jingsheng Zhai
Opt. Express 30(13) 22512-22522 (2022)

Processing of chromatic information in a deep convolutional neural network

Alban Flachot and Karl R. Gegenfurtner
J. Opt. Soc. Am. A 35(4) B334-B346 (2018)

High-quality color image restoration from a disturbed graded-index imaging system by deep neural networks

Xuanyu Hu, Zaipeng Duan, Yang Yang, Yehong Tan, Ruiqi Zhou, Jiong Xiao, Jinwei Zeng, and Jian Wang
Opt. Express 31(13) 20616-20628 (2023)