Abstract:

BACKGROUND: With the development of science and technology, there are more and more objective methods to research fatigue, and physiological indexes intervene makes it a medical, cognitive science and psychology research hotspot. However, the detection of mental fatigue is still lack of objective physiological indexes.
OBJECTIVE: To propose a novel sub-health state recognition method based on principal component analysis and pulse signal features is presented in order to evaluate mental fatigue state.
METHODS: Firstly, wavelet transform was used to de-noise the pulse signal. Secondly, the peak value of power spectrum, features of corresponding frequency, center of gravity and gravity frequency of power spectrum were extracted. Thirdly, the extracted features were performed with principal component analysis. Finally, an improved linear discriminant analysis was applied to classification and pattern recognition. The results demonstrated that the recognition ratio of the principal component was up to 100%.
RESULTS AND CONCLUSION: This method possessed many attractive characters such as simpler calculated process, better stability, and could get higher recognition rate, which provides a certain reference value for achieving detection of mental fatigue state.