Machine Intelligence based Detection and Classification of Human Physiology and Emotions

Abstract

Automated analysis of the physiological signals like ECG (electrocardiogram) and EEG (electroencephalogram) has become more extensive during the last three decades and is recognized as an effective clinical diagnostic tool in the physiological measurement field. Since most of the physiological signals including ECG and EEG are non-linear, non-stationary and non-Gaussian in nature, it is difficult to capture the respective large temporal and morphological variations for their automated analysis. Conventional physiological signal analysis tools such as linear and power spectrum estimation ignore random variations and the Fourier phase relationship among signal components. This can provide inaccurate analysis results. An effort shall be attempted to develop a more efficient and robust technique to analyze physiological signals using a fusion of time domain and frequency domain techniques. In this research, the physiological signals to be considered are ECG and EEG. The primary aim will be the analysis of ECG signals to develop an automated cardiac state detection and classification technique using machine intelligence tools like neural network classifier. The understanding obtained from primary aim shall be used to analyze EEG signals to interpret and classify correlated emotional states as this can lead to an effective way of implementing implicit man-machine interaction. newline