Abstract: An image restoration method, apparatus and device, and a storage medium are provided. The method comprises: evaluating, by an unsharpness evaluation network, an image to be processed, to generate an a feature map; inputting said image to be processed and the unsharpness feature map into a blur elimination main network; and iteratively processing, by the blur elimination main network, the image to be processed for a set number of times according to the unsharpness feature map to output a restored image. In the above-mentioned technical solution, an unsharpness feature of an image to be processed is extracted by using an unsharpness evaluation network, and iterative processing is performed on said image to be processed by using a blur elimination main network, where the networks are trained networks having corresponding functions, such that automatic restoration of an image can be realized, and image restoration efficiency and quality are improved.
Type:
Application
Filed:
December 29, 2021
Publication date:
February 22, 2024
Applicant:
HEFEI INNOVATION RESEARCH INSTITUTE OF BEIHANG UNIVERISTY