Method for extracting and classifying fMRI features based on adaptive entropy algorithm for projection clustering (APEC)

Abstract

The invention discloses a method for extracting and classifying fMRI features based on adaptive entropy algorithm for projection clustering (APEC). The method comprises the steps of building a dynamic function connecting matrix by the sliding time window method; performing (AEPC) to cluster dynamic function connecting matrixes of parties in A type and B type tests so as to form a plurality of clustering centers; calculating the similarity between the functional connecting matrix of each test party and each clustering center to form a similarity matrix; extracting elements in the similarity matrix as the features; classifying brain data through a training SVM classifier. With the adoption of the method, the generalization of a data classifying model is improved, and moreover, rich brain dynamic structure information can be extracted; the method can be applied to automatic handling and classification of brain data in the study of the biology information technology.

Description

technical field [0001] The present invention relates to an automatic processing and classification method of functional Magnetic Resonance Imaging (fMRI), in particular to a fMRI feature extraction and classification method based on an adaptive entropy projection clustering algorithm, which involves digital images Knowledge areas such as processing, dynamic functional connectivity matrix, Adaptive Entropy Algorithm for Projective Clustering (AEPC) and Support Vector Machine (SVM). technical background [0002] In recent years, with the continuous development of science and technology, the equipment and instruments for observing brain activity have been increasing, and the non-invasive method of measuring brain function activity has greatly promoted the development of brain neurofunction. Currently, the methods used to study brain activity mainly include magnetoencephalography (MEG), electroencephalography (EEG), single photon emission tomography (SPECT), electron emission to...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problem of fMRI data feature extraction and recognition, the present invention proposes a fMRI feature extraction and classification method based on an adaptive entropy projection clustering algorithm, which solves the problem of using traditional clustering for high-dimensional dynamic functional connectivity matrix data clustering recognition. The clustering effect of the class method is not good and the problem is affected by the subjective factors of the operator

More importantly, the present invention solves the problem that the traditional method cannot select the optimal clustering result, and can effectively obtain the optimal clustering center as the feature

Images