Animation drawing auxiliary creation method based on GAN

Abstract

The invention discloses an animation drawing auxiliary creation method based on a GAN. An integrated cartoon drawing auxiliary creation system is constructed by utilizing line draft professional enhancement technology based on generative adversarial network and reinforcement learning, line draft coloring technology based on related genus learning and color clustering, cartoon music album generation technology based on audio-visual association and emotion calculation and the like, so that the creation experience and artistic aesthetic appreciation of users are improved.

Description

technical field [0001] The present invention relates to the technical field of technical image generation, and more specifically, relates to a GAN-based animation drawing assistant creation method. Background technique [0002] At present, there are very rich image translation and style transfer tasks. Regardless of realistic or non-realistic rendering, it is to restore information or modify features of missing or to-be-transformed images. The deep learning network used in image generation is generally divided into variational autoencoder (VAE), generative confrontation network (GAN) and their variants. Because VAE has no confrontational learning, the generated pictures are blurred. The image generation of GAN and its variants is more widely used and the effect is better. [0003] Conditional GAN ​​adds some explicit external information to the original GAN, which is generally additional input from the user, such as semantic categories, high-dimensional categories, and und...

AI Technical Summary

Problems solved by technology

The coloring of the generated results of the first generation is relatively natural, but because vgg covers the topological space information corresponding to the color of the reference picture, the color structure of the picture to be colored is over-fitted, and it cannot be well-adapted to coloring

The third generation of PaintsTransfer is able to produce visually good shading results, but the local details of the color are dirty

[0013] Traditional animation creation requires professional painting skills and rich coloring experience, and creating high-quality animation images is time-consuming and labor-intensive

At present, there are some problems that need to be optimized in the existing animation creation system, such as the professional enhancement of the line draft, the unreasonable spread of the color of the line draft coloring, the low level of intelligence of the automatic coloring, the effect of interactive coloring and the lack of user experience. , Full-process and efficient generation of animation music albums, etc.

Images