Mammary gland X-ray image automatic generation method based on convolutional generative adversarial network

Abstract

The invention provides a mammary gland X-ray image automatic generation method based on a convolutional generative adversarial network, and the method comprises the following steps: 1, carrying out the preprocessing of an inputted original image, removing the redundant background in the image, and adjusting the proper size to enable the lump and mammary gland background of a lump region to be segmented; step 2, constructing a convolutional generative adversarial network model which comprises a generative network G, a discriminant network D and a pre-trained discriminant network D-pro; 3, pre-training the convolutional generative adversarial network, and storing trained model parameters as initialization parameters of the convolutional generative adversarial network; 4, generating an image;and step 5, fusing the breast lump image and the breast background generated in the step 4 based on medical features. According to the mammary gland X-ray image automatic generation method based on the convolutional generative adversarial network provided by the invention, lump images with different morphologies and sizes are generated by adopting a thought of firstly segmenting, then generatingand then fusing, and a multi-lump mammary gland image is generated based on the single-lump mammary gland images, so that data support is provided for research of a multi-lump mammary gland diagnosistechnology.

Description

technical field [0001] The invention relates to the technical field of computer deep learning, in particular to a method for automatically generating breast X-ray images based on a convolutional generative confrontation network. Background technique [0002] Breast cancer is one of the most common cancers in women, with a global incidence of 24.2% among female cancers and 13.7% of cancer-related mortality. The etiology and pathogenesis of breast cancer are complex and difficult to cure. Improving diagnostic techniques for screening and timely treatment can effectively increase the cure rate. Breast mammography screening is the most reliable method for early detection of breast cancer. In breast abnormalities such as microcalcifications and structural distortions, it is possible to confirm whether a breast mass is malignant or benign as early as possible through screening. The diagnosis of mammary X-ray images is mainly manual observation, classification and judgment, mostly...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a method for automatic generation of breast X-ray images based on convolutional generation confrontation network, to solve the problem that the features extracted based on the twin network contain the information of the template and the search area, and the target positions in them are constantly changing. The problem that the center position of the marked target may be inaccurate due to changes and slight differences in the extracted features

Images