Image translation method and device and image translation model training method and device

Abstract

The invention discloses an image translation method and device and an image translation model training method and device, and relates to the technical field of deep learning and image processing. According to the specific implementation scheme, the method includes obtaining an image translation request, and the translation request comprises an original image; processing the original image to generate a pre-translated image, a mask image and deformation parameters corresponding to the original image; performing deformation processing on the original image according to the deformation parametersto obtain a deformed image; and fusing the deformed image, the pre-translated image and the mask image to generate a target translated image. Therefore, the definition of the generated target translation image is improved, and the background part of the generated target translation image is consistent with that of the original image, so that seamless fusion of the images can be realized, and thenaturalness of the generated target translation image is greatly improved.

Description

technical field [0001] The present application relates to the technical field of image processing, specifically to the technical field of deep learning and image processing, and in particular to an image translation method and device, and an image translation model training method and device. Background technique [0002] The image translation network can directly convert one type of image into another type of image without changing the image content, and has a wide range of applications in image generation, scene segmentation, image stylization and other fields. [0003] In related technologies, traditional image translation methods, for example, an image conversion method based on CycleGAN, are generally used to translate images. However, the traditional image translation method can only directly convert one type of image into another type of image, the definition of the converted image is low, and the background part of the converted image also exists with the original im...

AI Technical Summary

Problems solved by technology

However, the traditional image translation method can only directly convert one type of image into another type of image, the definition of the converted image is low, and the background part of the converted image also exists with the original image. Differences, resulting in obvious borders when the images are fused, and it is impossible to achieve natural convergence and seamless fusion

Images