Semi-supervised cervical cancer image classification method based on virtual adversarial training network

The invention discloses a semi-supervised cervical cancer image classification method based on a virtual adversarial training network, and the method comprises the steps: collecting a cervical cancer image data set, and obtaining a training data set and a test data set for a semi-supervised network; carrying out data enhancement on the labeled and unlabeled cervical cancer images in the training data set, wherein translation transformation, rotation transformation and scaling transformation are carried out on the cervical cancer images; building a backbone network, wherein the backbone network comprises AlexNet and a mixed attention mechanism; a virtual adversarial training network is built, given input is adopted, disturbance is increased, and the model gives different outputs; the model punishes the sensitivity due to disturbance; related features of the cervical cancer image are obtained through the virtual adversarial training network and sent to the classifier for classification, and finally the category