Image processing network training method, computer equipment and readable storage medium

Abstract

The invention relates to an image processing network training method, computer equipment and a readable storage medium. The method comprises the following steps: acquiring a plurality of training sample images, wherein the plurality of training sample images are a plurality of groups of tomographic images; inputting the plurality of training sample images into an initial image processing network for image processing to obtain image processing results of the plurality of training sample images; and calculating the loss between the image processing results of the plurality of training sample images and the corresponding gold standards, calculating the difference value between the image processing results of the adjacent layers of training sample images in the same group of training sample images, and training the initial image processing network according to the loss and the difference value. According to the method, when the loss of the image processing result of the training sample image is calculated, the difference value between the image processing results of the adjacent layers of training sample images is considered, and meanwhile, the number of the training sample images is not reduced, so that the network precision obtained by training is greatly improved.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to an image processing network training method, computer equipment and a readable storage medium. Background technique [0002] In the medical field, it is usually necessary to collect medical images of patients for further detection and diagnosis by doctors. At present, the commonly used imaging technologies include computerized tomography (Computed Tomography, CT), nuclear magnetic resonance imaging (Nuclear Magnetic Resonance Imaging, MRI), positron emission computed tomography (Positron Emission Computed Tomography, PET), etc., the medical images obtained by these imaging techniques are usually multi-layer two-dimensional images. Taking the detection of hemorrhage in brain images as an example, the doctor's test results directly affect the patient's follow-up treatment, so the accuracy of the test results is particularly important. [0003] Therefore, the dee...

AI Technical Summary

Problems solved by technology

[0004] Based on this, it is necessary to provide a training method, computer equipment and readable storage medium for an image processing network to solve the problem of low network accuracy obtained by traditional techniques for deep learning network training.

Images