| abstract |
The invention discloses a blind face restoration method and system, comprising: collecting blind face data sets, using Laplace gradient to evaluate the quality of the blind face data sets, and removing blurred and non-face images; The AFFNet network is constructed by randomly assigning the training set and test set to enhance the image data, and the image of the training set is input into the AFFNet network, and the joint reconstruction loss function, perceptual loss function, style loss function and adversarial loss function are used for the AFFNet network. Carry out training, and use the SGD optimization algorithm to train and optimize the AFFNet network to obtain the optimal blind face restoration model; input the images of the test set into the optimal blind face restoration model, and perform matching and selection to obtain the image with the highest accuracy as the result of the final search. Through the above solution, the present invention has the advantages of simple logic, accuracy and reliability, etc., and has high practical value and promotion value in the field of image processing technology. |