The invention belongs to the technical field of medical
signal processing, and particularly relates to a
coronary heart disease detection
system for heart sound and electrocardio
correlation analysis, which can perform data preprocessing, including
noise reduction,
baseline drift elimination,
power line interference and the like, on PCG and ECG original signals, perform normalization
processing on the PCG and ECG signals, and improve the accuracy of
data processing. Performing
wavelet transform on the processed heart sound and electrocardiosignals to generate a two-dimensional time-frequency image, performing
feature extraction on the processed signals and the time-frequency image by using a 1D CNN and a 1D CNN, performing feature splicing on 1D CNN features and 2D CNN features in the same mode through global
pooling, generating fusion features with
discriminant lines by using a
discriminant correlation analysis method, and performing
feature fusion on the fusion features with
discriminant lines. Performing multi-classification on the diseases through a
support vector machine; in order to effectively utilize the correlation between different
modal signals, a discriminant
correlation analysis method is adopted to maximize the correlation between
modal features, eliminate the inter-class correlation and retain the intra-class correlation, the inter-
modal correlation features are enhanced, and multi-classification of the
coronary heart disease is realized.