Image classification network training method and device, equipment and storage medium

Abstract

The invention discloses an image classification network training method, device and equipment, and a storage medium. The method comprises the steps of generating an image classification network basedon a labeled sample set; selecting target data from the unlabeled data for labeling based on the strategy network and the evaluation network; updating the labeled sample set according to a labeling result of the target data, and updating the image classification network according to the updated labeled sample set; updating the strategy network and the evaluation network according to the labeling result; and circularly executing the selection operation and the updating operation until a preset condition is met to obtain a trained image classification network. According to the technical scheme,only part of unlabeled data is selected each time for labeling and updating the image classification network, so that the data cost and labeling workload in the training process are reduced, and the classification precision is ensured.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of image recognition, and in particular to a training method, device, equipment and storage medium for an image classification network. Background technique [0002] With the rapid development of medical imaging equipment and artificial intelligence technology, medical data has grown on a large scale, and the analysis of artificial intelligence medical images based on deep learning has been gradually applied. An image classification network can be trained based on a large amount of labeled data, and the application of the image classification network can identify specific features, thereby identifying and classifying images, greatly reducing the workload of doctors, and improving diagnostic efficiency and accuracy. [0003] However, for the image classification network, a large amount of labeled data is required to train the model, which requires a lot of manual labeling work by pro...

AI Technical Summary

Problems solved by technology

[0003] However, for the image classification network, a large amount of labeled data is required to train the model, which requires a lot of manual labeling work by professional doctors, and in the medical field, it is also difficult to obtain a large amount of medical image data, which makes the training process relatively difficult. Complex and expensive; in addition, although a large amount of training data is necessary for the training of deep neural networks, not all of the data are necessary. If the training data contains noise or wrongly labeled data, it will make the neural network The classification accuracy decreases. Therefore, training based on a large amount of or unnecessary data affects the training efficiency and classification accuracy of the image classification network, and the cost is high and the implementation is complicated.

Images