Generative confrontation network training method, image completion method, equipment and storage medium

Abstract

Embodiments of the present application provide a generative adversarial network training method, an image completion method, a device, and a storage medium. In some exemplary embodiments of the present application, first, image completion training is performed using sample images containing missing regions to obtain a preliminary completion network and preliminary completion images; then, discriminator training is performed using preliminary completion images, The first local context discriminator, the second local context discriminator and the global discriminator are obtained; finally, using the sample image containing the missing area, the primary The completion network conducts confrontational training to obtain an image completion network; the first local context discriminator maintains the local consistency of image completion, the global discriminator maintains the global consistency of image completion, and the second local context discriminator ensures the consistency of texture information. Authenticity and consistency with complementary central and surrounding areas.

Description

technical field [0001] The present application relates to the technical field of data processing, and in particular to a generative adversarial network training method, an image completion method, a device, and a storage medium. Background technique [0002] Image completion techniques aim to synthesize missing or damaged regions in images, which is a fundamental problem in low-level vision. The technique has attracted widespread interest in computer vision and graphics because it can be used to fill in occluded image regions or repair damaged photos. In addition, before sharing a photo, users may need to make modifications to the image, such as erasing distracting scene elements, adjusting the position of objects in the image for better composition, or restoring image content in occluded image areas, etc. These and many other editing operations require automatic completion of missing regions in images, which has been an active research topic in the computer vision and grap...

AI Technical Summary

Problems solved by technology

Therefore, the existing results of image completion using GAN methods cannot maintain the consistency of image spatial structure and context information

Images