Face age change image confrontation generation method and system

Abstract

The invention relates to a face age change image confrontation generation method and system. The generation method comprises the steps of acquiring multiple pairs of real face images and target age feature vectors; enabling a face generator based on an airspace attention mechanism to obtain a composite image according to each pair of real face images and the target age feature vector; based on a face discriminator, calculating a loss value of an image loss function according to each real face image and the corresponding composite image; according to the loss value, iteratively adjusting the weights of a face generator and a face discriminator by using a loss gradient back propagation algorithm until convergence; and based on the current face generator, according to the to-be-processed faceimage and the corresponding target age feature vector, obtaining a face image with the target age feature. According to the invention, the change area of the input image passing through the generatoris limited through the spatial domain attention mechanism, so that the possibility of pixel change in the image area irrelevant to age change can be reduced, and the probability of introducing noiseand distortion can be reduced.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a method and system for confrontation generation of face age change images. Background technique [0002] Synthesizing face images after age changes is an important branch of image editing in the field of computer vision. Face images with age-specific appearance characteristics. [0003] With the wide application of deep learning theory, especially the rapid development of Generative Adversarial Networks (GANs), most of the existing face age change technologies use GANs-based models to synthesize realistic face images. Generally speaking, age changes only focus on local areas of the face (forehead, eye corners, mouth, etc.), and a reasonable age change model should only focus on pixel changes in these areas, while keeping other parts of the image information unchanged. [0004] However, most existing models change the entire input image, resulting in changes in...

AI Technical Summary

Problems solved by technology

[0004] However, most existing models change the entire input image, causing the pixel values ​​in the image background and other areas that are not related to age changes to change, thereby introducing noise and distortion, reducing the quality and fidelity of the generated image.

Images