ANFIS-GA system for three-dimensional pulse image of normal and string-like pulse in Chinese medicine using an improved contour analysis method

Introduction: In Chinese Medicine (CM), pulse feeling has been translated into a three-dimensional pulse image (3DPI), a visible quantitative format. An adaptive neuro-fuzzy inference system (ANFIS) with a genetic algorithm (GA) with several spatiotemporal indexes from contour analysis of 3DPIs was proposed in order to investigate the taut level of a string-like pulse. Methods: Forty-eight pulse image patterns and 283 2DPIs were obtained before and after a cold pressor test (CPT). The 3DPI and two-dimensional pulse image (2DPI) were constructed successively via spline interpolation. The improved contour analysis of pulse images was to elucidate three indexes from 2DPI at the contour level 0.8: the longitudinal diameter (2DL), the transverse diameter (2DT), and the area of contours. These indexes were adapted to ANFIS-GA system, whereas the output was taut level value α. Results: α can reflect the taut level of string-like pulse. The ANFIS-GA provides a pattern and system to analyze taut level of string-like pulse. Conclusions: CPT increases arterial stiffness to make pulse image tauter. The ANFIS-GA can efficiently demonstrate the taut level of pulse images to help CM physicians diagnose CM syndromes and promote CM standardization.