Training method of image generation network, image prediction method and computer device

Abstract

The invention relates to a training method of an image generation network, an image prediction method and a computer device. The training method comprises the steps of obtaining a training sample image; inputting the training sample image into an initial image generation network to obtain an initial prediction image; wherein the initial prediction image comprises a preset simulation mark, and theinitial prediction image represents the change condition of the predicted measured object after a preset time interval; inputting the initial prediction image and the first real sample image into an initial discrimination network to obtain a discrimination result of the initial prediction image; wherein the first real sample image is a reference image of the initial prediction image; calculating the loss between the discrimination result and the simulation mark, and training the initial discrimination network and the initial image generation network according to the loss; and when the loss reaches convergence, completing training of the initial image generation network to obtain an image generation network. According to the method, the precision of the image generation network obtained bytraining can be improved, and the accuracy and intuition of the obtained prediction image can be improved.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to an image generation network training method, image prediction method and computer equipment. Background technique [0002] In the medical field, it is usually necessary to predict the future state of the patient, such as whether the current hematoma will expand over time, whether the current pulmonary nodule will transform, etc. Similar prediction tasks can be collectively referred to as longitudinal change prediction. In terms of clinical application, the high-precision longitudinal change prediction of scan data can provide doctors with intuitive and accurate longitudinal data information, and it is also widely used in drug effect tracking and patient condition return visits, so it has great potential in computer-aided diagnosis and clinical diagnosis. important role. Magnetic resonance imaging is widely used in the diagnosis of brain diseases because of its...

AI Technical Summary

Problems solved by technology

[0005] Based on this, it is necessary to provide a training method for an image generation network, an image prediction method and a computer device for the low accuracy and poor intuitiveness of the vertical change prediction method in the traditional technology

Images