A medical image segmentation method, system and electronic device based on generative confrontation network

Abstract

The present application relates to a medical image segmentation method, system and electronic equipment based on generative confrontation network. First, study how the generator extracts the pixel-level features of different categories of high-quality images, and use the capsule model for structured feature representation, and then realize the generation of pixel-level labeled samples; secondly, build a suitable discriminator for discriminant generation Mark the authenticity of the sample at the pixel level, and design an appropriate error optimization function, and feed back the discrimination results to the models of the generator and the discriminator respectively. Through continuous confrontation training, the sample generation capabilities and discrimination ability, and finally use the trained generator to generate pixel-level labeled samples to achieve pixel-level segmentation of image-level labeled medical images. The present application effectively reduces the dependence of the segmentation model on pixel-level labeled data, not only improves the efficiency of confrontation training between generated samples and real samples, but also effectively realizes high-precision pixel-level image segmentation.

Description

technical field [0001] The present application belongs to the technical field of medical image processing, and in particular relates to a medical image segmentation method, system and electronic equipment based on generative adversarial networks. Background technique [0002] With the vigorous development of medical imaging technology, medical imaging has been widely and deeply applied in clinical medicine. According to statistics, tens of millions of cases are diagnosed and treated through medical imaging every year around the world. In the traditional method of diagnosis and treatment based on medical images, doctors read and recognize medical image data, and make judgments on the diagnosis and treatment of diseases. This method of diagnosis and treatment is very inefficient, and there are large individual differences. Doctors can easily miss and misdiagnose based on their personal experience. Long-term reading of films will lead to fatigue of doctors and a decrease in th...

AI Technical Summary

Problems solved by technology

The existing image segmentation model based on generative adversarial networks can be applied to cross-category object image segmentation, but in the field of medical images, the model has problems such as insufficient feature extraction and a large amount of calculation for adversarial training.

Images