Alzheimer's disease classifier based on brain imaging big data deep learning

Abstract

The invention discloses an Alzheimer's disease classifier based on brain imaging big data deep learning. The Alzheimer's disease classifier comprises a data preprocessing module, a deep learning model, a gender classification module, an initialization module, an AD training module and a prediction module. The grey matter density and grey matter volume images registered by the data preprocessing module are input by the input module, and image feature values are extracted after convolution, reduction and pooling processing; carrying out gender deep model training on the brain imaging big database sample through a deep learning model; when the gender classification accuracy reaches the highest value; and then parameter initialization is performed on a drop out module and a Softmax function inthe deep learning model through an initialization module, finally AD training is performed on a big database sample through an AD training module by using the deep learning model, and AD detection and classification are performed after training. According to the method, the classification accuracy of AD patients and normal people is remarkably improved, the AD classification accuracy reaches 88.4%, and the AD classification accuracy on independent samples reaches 86.1%.

Description

technical field [0001] The invention relates to the technical field of brain imaging, in particular to an Alzheimer's disease classifier (Alzheimer's Disease, AD) based on deep learning of brain imaging big data. Background technique [0002] Brain imaging, especially magnetic resonance imaging, has been developed for decades, but its clinical application is still relatively limited. However, big data and deep learning have made breakthroughs in imaging-based diagnosis of breast cancer and fundus diseases. At present, due to the high cost of magnetic resonance imaging and high requirements for patient cooperation, the accumulation of big data and related research on magnetic resonance has been on the order of thousands of people, and it is difficult to construct an industrial-grade deep learning classifier based on big data. For practical purposes, its universality is poor. Contents of the invention [0003] In order to solve the above-mentioned technical problems, based...

AI Technical Summary

Problems solved by technology

At present, due to the high cost of magnetic resonance imaging and high requirements for patient cooperation, the accumulation of big data and related research on magnetic resonance

Images