Feature extraction method for automatic recognition of identity through electrocardiosignal

Abstract

The invention discloses a feature extraction method for the automatic recognition of identity through an electrocardiosignal, and the method comprises the steps: (1), enabling a training set electrocardiosignal to be converted into a training set pseudo cardiac beat signal, removing the noise in the training set electrocardiosignal, searching the second-order differential zero point of the training set electrocardiosignal, detecting an R wave of the training set electrocardiosignal, and forming a training set pseudo cardiac beat signal matrix through segmenting; (2), extracting the features of the training set electrocardiosignal through employing the training set pseudo cardiac beat signal matrix: calculating the power spectrum density matrix and autocorrelation matrix of the training set pseudo cardiac beat signal matrix, and carrying out the fusion of training set transform domain features and training set time domain features; (3), converting a to-be-recognized electrocardiosignal into a to-be-recognized pseudo cardiac beat signal; (4), extracting the features of the to-be-recognized electrocardiosignal through the to-be-recognized pseudo cardiac beat signal. The method effectively reduces the signal collection time needed by the identity recognition, improves the recognition accuracy, and reduces the calculation cost.

Description

technical field [0001] The invention relates to the technical field of identification, in particular to a feature extraction method for automatic identification using electrocardiographic signals. Background technique [0002] The electrocardiogram is unique to an individual and can be used as an important basis for human identification. The identification device based on the electrocardiographic signal refers to a device that can automatically identify the identity of the person who receives the electrocardiographic signal from outside. [0003] The definition of using ECG signal for identity recognition is: given an ECG signal, determine the identity of the person to whom the signal belongs. A system or device for automatic identification using ECG signals is usually implemented based on the pattern recognition technology of ECG signals, and two necessary links, feature extraction and recognition of ECG signals, need to be realized. Among them, the feature setting in the...

AI Technical Summary

Problems solved by technology

[0004] In the disclosed inventions of devices and methods for identification using ECG signals, features are often extracted through a relatively long period of ECG signals, and there are two problems. On the one hand, users need to wait for a long time to perform ECG signals On the other hand, a larger amount of data requires more computing costs

Images