Design of an Intelligent Identity Detection Systembased on ECG Signal Processing

The main purpose of this paper is to develop a method to investigate heart signals for detection of the individuals. In this regard, data sets of 48 dual-channel ECG records were received from MIT-BIT data banks and used as the samples. Each of these samples was extracted during one hour at 128 Hz sampling rate for each individual. Each file was divided to time intervals of 10 seconds long and then analyzed using a wavelet transform down to level 10. For every level, entropy, power, energy, variance, standard deviation and average values (totally 60 features) were derived. All these 60 features were introduced for the first time, with the responding time and percentage being compared with those selected by an IDE feature selection. For this purpose, a threshold limit of the IDE equal to 0.8 was considered. Therefore, out of the 60 suggested features, only some 7 features (i.e. standard deviation of signal, energy of signal, variance of signal, power of signal, standard deviation of signal, power of signal, and standard deviation of signal) were found to be below the threshold limit. Afterwards, these 7 features were applied by difference classification methods to get the responding percentage and computational cost. The computational cost of knn method is smaller than other methods in both cases of selecting the features. On the other hand, in terms of accuracy of the XCSLA, selection of the features is the best method which has detected all classifications correctly using the IDE algorithm. Taking into account the results obtained from application of the developed classification systems for detecting identity of the individuals from their ECG signals based on their clinical symptoms can lead to make this conclusion that the proposed system may be an appropriate alternative for detection of the individuals as a biometric method. However, it is still necessary to conduct further research works on such a system. The main drawback of this system could stem from different feelings of a human being including anger, happiness, sadness and etc. which may change rhyme of the heartthrob

کلیدواژه ها:

Biometric methods ، ECG ، XCSLA