Magnetic resonance image feature extraction and classification method based on deep learning

Abstract

The invention provides a magnetic resonance image feature extraction and classification method based on deep learning, comprising: S1, taking a magnetic resonance image, and performing pretreatment operation and feature mapping operation on the magnetic resonance image; S2, constructing a multilayer convolutional neural network including an input layer, a plurality of convolutional layers, at least one pooling layer / lower sampling layer and a fully connected layer, wherein the convolutional layers and the pooling layer / lower sampling layer are successively alternatively arranged between the input layer and the fully connected layer, and the convolutional layers are one more than the pooling layer / lower sampling layer; S3, employing the multilayer convolutional neural network constructed in Step 2 to extract features of the magnetic resonance image; and S4, inputting feature vectors outputted in Step 3 into a Softmax classifier, and determining the disease attribute of the magnetic resonance image. The magnetic resonance image feature extraction and classification method can automatically obtain highly distinguishable features / feature combinations based on the nonlinear mapping of the multilayer convolutional neural network, and continuously optimize a network structure to obtain better classification effects.

Description

technical field [0001] The invention relates to the technical field of medical data processing, in particular to a magnetic resonance image feature extraction and classification method. Background technique [0002] Magnetic resonance imaging is a medical imaging technique with high resolution, no damage, and clear anatomical structures. Therefore, magnetic resonance imaging plays an important role in the medical diagnosis process, especially in the diagnosis and research of brain tissue-related diseases. a wide range of applications. In the past decade, Magnetic Resonance Imaging (MRI, Magnetic Resonance Imaging), as an important brain imaging method in neuroimaging, has been widely used in the diagnosis of neurological and mental diseases without brain tissue / structural damage. in the study of neuroscience and brain disorders. [0003] Machine learning algorithms such as MVPA (Multi-voxel Pattern Analysis) are often used to identify spatial patterns in human brain magnet...

AI Technical Summary

Problems solved by technology

[0005] In view of the above problems, the present invention aims to provide a deep learning-based magnetic resonance image feature extraction and classification method, which effectively solves the problem of low classification accuracy due to the need for manual selection of feature input in the traditional magnetic resonance image classification process. The non-linear mapping of multi-layer convolutional neural network can automatically obtain highly separable features / feature combinations for classification, and can continuously optimize the network structure to obtain better classification results

Images