Prostate Multimodal MR Image Classification Method and System Based on Fovea Residual Network

Abstract

The invention discloses a prostate multimode MR image classification method and system based on a fovea residual network, wherein the method includes: using the fuzzy kernel in the fovea operator to replace the convolution kernel of the residual network, thereby constructing the fovea Residual network; use the multimodal MR images of the prostate with class labels to train the fovea residual network to obtain the trained fovea residual network; use the fovea residual network to classify the prostate multimodal MR images to be classified, and obtain the classification result. Based on the visual characteristics of the human eye, the present invention designs a fovea operator, extracts the fuzzy kernel of the operator, and uses it to replace the convolution kernel in the residual network, thereby constructing a fovea deep learning network, which can extract features, thereby improving the classification accuracy of prostate multimodal MR images.

Description

technical field [0001] The invention belongs to the field of image classification, and more particularly, relates to a method and system for classifying prostate multimodal MR images based on a foveal residual network. Background technique [0002] Prostate cancer is the most common malignant tumor in the male reproductive system. According to the American Cancer Society report in 2018, prostate cancer ranks first in new cases and second in deaths in men. In recent years, with the significant improvement of people's living standards and changes in the age structure of the population, the incidence and mortality of prostate cancer in my country have shown a significant upward trend. [0003] MRI can reflect the changes of tissue and internal organ functions and biochemical metabolic reactions, and has the advantages of high-resolution imaging of soft tissue, and has important application value in the early diagnosis of prostate cancer. In recent years, more and more new MR i...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the above defects or improvement needs of the prior art, the present invention provides a method and system for the classification of prostate multimodal MR images based on the fovea residual network, thereby solving the problems in the prior art that it is difficult to extract features that conform to the visual characteristics of the human eye , the technical problem of low accuracy of prostate image classification based on deep learning

Images