Image processing method and system based on deep neural network, and medium

Abstract

The invention provides an image processing method and system based on a deep neural network, and a medium. The image processing is to restore a motion blurred image. The image processing method comprises the following steps: S1, establishing a sample database which stores images with motion blur and comprises a training set and a test set; S2, constructing and optimizing the deep neural network used for restoring the motion blurred image based on a twinning network framework of a ResNet residual network by using the training set; and S3, restoring the motion blurred image in the test set by using the deep neural network.

Description

technical field [0001] The present invention relates to the field of image processing, in particular to an image processing method, system and medium based on a deep neural network. Background technique [0002] Image restoration technology aims to restore degraded low-quality images, so as to obtain more effective information. The existing restoration methods have certain limitations in the restoration of motion blurred images, such as slow restoration speed and low restoration definition. When collecting clear images at ordinary times, there are many factors that may cause image blurring. There is a certain amount of relative motion between the camera lens and the captured scene, resulting in blurred images. [0003] Image restoration technology aims to repair low-quality images caused by adverse environmental factors in the shooting process, and try its best to make low-quality images clear, so as to present more effective information. The restoration technology of mot...

AI Technical Summary

Problems solved by technology

Jain et al. used a 4-layer CNN model to restore blurred images, but the experimental results show that CNN with shallower network layers is still not very effective in image restoration.

Images