Semi-supervised learning image segmentation method, system and terminal

Abstract

The invention discloses a semi-supervised learning image segmentation method and system, and a terminal. The method comprises the following steps: training an image segmentation network by using a first data set of an image with a label; inputting the data set of the non-label image into the trained image segmentation network to obtain a pseudo-label data set corresponding to the non-label image; combining the pseudo-label data set and the initial data set with the tag image to obtain a second data set; and training the image segmentation network again by adopting the second data set, inputting the next batch of label-free image data set after training is completed, and predicting to generate false labels. After a pseudo tag is generated and expanded to a training set every time, due to the fact that new training data is added, the network continuously learns new features until a trained segmentation network is finally obtained. According to the method, a small amount of marked data and a large amount of unmarked data are jointly used to effectively train the segmentation network, so that the purpose of improving the tissue segmentation precision is achieved, and the dependence of a deep learning image segmentation method on label data is reduced.

Description

technical field [0001] The present application relates to the technical field of image segmentation processing, in particular to an image segmentation method, system and terminal for semi-supervised learning. Background technique [0002] In recent years, deep learning methods represented by deep convolutional neural networks can automatically extract a large number of effective high-level features by learning a large number of labeled samples, thereby improving the accuracy of tissue segmentation. Fully convolutional neural networks can directly process the entire image, enabling end-to-end image segmentation. In 2015, Ronneberger proposed UNet and applied it to the field of biomedical image segmentation, which is the first application in the field of biomedical image segmentation. Since UNet can combine high-level semantic information and low-level information, many researchers at home and abroad have applied UNet as a backbone network to many automatic tissue segmentatio...

AI Technical Summary

Problems solved by technology

In the training data of the neural network, only a small number of images are labeled, while a large number of images have no corresponding labels. However, image labeling is a time-consuming and labor-intensive process.

Images