ST SEGMENT CLASSIFICATION CONVOLUTIONAL NEURAL NETWORK BASED ON FEATURE SELECTION AND METHOD FOR USING SAME

摘要

An ST segment classification convolutional neural network based on feature selection, and a method for using same. A plurality of convolutional neural networks of different structures are used, some output with higher relevancy is screened out from output of the last pooling layer of the plurality of convolutional neural networks to serve as input of an independent full connection layer, and finally the classification type of an ST segment is determined according to output of the independent full connection layer.