Processing method of image generator and image generation method and device

Abstract

The invention relates to a processing method of an image generator and an image generation method and device. The processing method of the image generator comprises the following steps: acquiring a source domain image sample and a reference image sample; generating a target generation image of the source domain image sample in the target domain through an image generator; respectively extracting afirst content feature of the source domain image sample, a second content feature of the reference image sample and a third content feature of the target generation image; generating a positive sample according to the first content feature and the third content feature, and generating a negative sample according to at least one of the first content feature and the third content feature and the second content feature; and inputting the positive sample and the negative sample into a mutual information discriminator, carrying out iterative adversarial training on the image generator and the mutual information discriminator, and iteratively maximizing the mutual information of the first content feature and the third content feature in the adversarial training process until an iterative stop condition is reached. By adopting the method, deformation of the migrated image can be avoided.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a processing method of an image generator, an image generating method and a device. Background technique [0002] With the development of artificial intelligence, the use of machine learning models is becoming more and more widespread. By training the machine learning model, the machine learning model performs data processing. For example, the medical image processing model is trained so that the medical image processing model can realize the classification or segmentation of medical images. [0003] Currently, when training a machine learning model with image samples, the image samples may come from different image domains. For example, the image samples in the training set are darker and the image samples in the test set are brighter, which will affect the performance of the machine learning model. In the traditional technology, the image samples can be transferre...

AI Technical Summary

Problems solved by technology

However, traditional image migration methods can lead to distortion of the transferred image

Images