Digital image restoration method and device based on edge information guidance

Abstract

The invention discloses a digital image restoration method and device based on edge information guidance, and the method comprises the steps: constructing an adversarial neural network which is composed of an edge prediction network and a content generation network; acquiring an original image data set with the same category as the to-be-restored digital image, randomly cutting the original image data set, adding random masks in batches to generate a stained image data set, and taking the stained image data set as a training data set; and inputting the training data set into the adversarial neural network, and training the edge prediction network and the content generation network respectively to obtain a trained edge prediction network and a trained content generation network. The edge information of the image is adopted to guide the neural network to repair the digital image, and the introduction of the edge information is beneficial for the neural network to better predict the structure of the stained region of the image, so that the repair capability of the neural network for any large-area stained region is improved.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a digital image restoration method and device guided by edge information. Background technique [0002] A digital image is a two-dimensional image represented by pixels in a limited numerical form, and is composed of a matrix that stores the specific numerical information of each pixel in a computer. There are often such undesirable areas in digital images: scans of originally scratched or defaced photos; unwanted target objects or text appearing on images; The missing information area left by the loss, etc. In order to ensure the integrity of image information and visual consistency, digital image restoration techniques are usually used to repair similar unideal areas. [0003] Image inpainting technology refers to the technical means of inferring and repairing the missing part of the information with the help of known image information, or repairing the defaced area...

AI Technical Summary

Problems solved by technology

However, the above-mentioned method based on deep learning is still based on the traditional RGB three-channel, and the image features are learned through the convolutional neural network, and then the damaged area is repaired. The MASK mask introduced is only used to mark the damaged area, which cannot give the neural network an effective Guided, blurring and artifacts tend to appear in the face of large areas of damage

Images