Picture classification method and device based on semi-supervised learning and computer equipment

Abstract

The invention discloses a picture classification method and device based on semi-supervised learning, computer equipment and a storage medium. The method comprises the steps that acquiring OCT pictures to be classified; processing the OCT pictures to be classified by using a feature vector generator in a preset OCT picture classification model to obtain a first feature vector X generated by a first encoder; decoding the first feature vector X by using the first decoder to obtain a decoded picture; generating a second feature vector Y by using the second encoder; calculating a similarity valueof the first feature vector X and the second feature vector Y according to a preset similarity calculation method, and judging whether the similarity value is greater than a preset similarity threshold value or not; and if the similarity value is greater than the preset similarity threshold, classifying the OCT pictures to be classified as negative pictures. Therefore, OCT picture classification is completed without positive data, and the defect that the positive data is difficult to collect is overcome.

Description

technical field [0001] The present application relates to the computer field, in particular to a semi-supervised learning-based picture classification method, device, computer equipment and storage medium. Background technique [0002] OCT (optical coherence tomography, optical coherence tomography) images have become a relatively common clinical ophthalmic disease examination and diagnosis method, and it is of great significance to screen ophthalmic diseases based on OCT images based on artificial intelligence methods. A common solution is to define this type of problem as a binary classification task, by collecting a batch of normal negative samples and abnormal positive samples, that is, each OCT image is given a label, and then choose a different classification model, After training with a certain amount of data, it can automatically predict the categories of normal and abnormal images, and realize ophthalmic disease screening based on OCT images. However, this type of ...

AI Technical Summary

Problems solved by technology

However, there are two obvious defects in this kind of supervised learning method: (1) It needs to collect a large number of samples for model training, especially abnormal pictures

In the field of medical images, it is relatively easy to collect normal negative samples, while the collection of abnormal samples is time-consuming and needs to be marked by experts, which is costly; (2) Due to the difficulty of collecting positive samples, it is very easy to cause unbalanced sample distribution, that is, normal There are many samples and few abnormal samples, and it is difficult to guarantee the accuracy of the trained model. Especially for disease screening scenarios, there is a great risk in the recall rate, which needs to be improved

Therefore, traditional techniques cannot accurately classify OCT images without obtaining a sufficient number of positive samples

Images