Evaluation method of physiological state by independent component analysis applied to facial thermal image

There have been various investigations of image recognition and assessment of human physiological state by infrared thermography. Assessing physiological state by infrared thermography exploits the nasal and forehead skin parts of facial skin temperature information, whereas the other parts are much less used. The objective of the present study was to propose a physiological measurement of a facial thermal image (FTI) by independent component analysis (ICA). ICA is a well-known statistical analysis method that estimates an original source signal from observed mixture signals. In this study, the authors attempted to extract a physiological index for an FTI by applying ICA. The extracted independent component represented temperature fluctuations from the opening and closing of the eyes and respiration, and the results obtained suggest that the proposed method is able to extract a physiological index.