Pressure wrist pulse signal analysis by sparse decomposition using improved Gabor function.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Medicine

Background:

• Wrist pulse wave analysis is crucial in traditional Chinese and Ayurvedic medicine for health assessment.
• Modern sensing and computational methods are increasingly used for analyzing pulse wave signals.
• Quantifying pulse wave peaks is significant for identifying health status.

Purpose of the Study:

• To develop a computational method for analyzing radial artery pressure pulse waveforms.
• To improve the representation and quantification of pulse wave peaks for disease diagnosis.
• To evaluate the effectiveness of the proposed method in distinguishing healthy and diseased pulse signals.

Main Methods:

• Decomposition of radial artery pressure pulse waveforms using sparse decomposition with an improved Gabor function.
• Design of an improved Gabor function structure for time-frequency dictionary generation to represent peak characteristics.
• Utilizing feature vectors from decomposed components for computerized pulse signal analysis.

Main Results:

• Achieved high accuracies in binary classification: 93.54% (health/diabetes), 73.42% (health/cardiac disease), 88.42% (health/hypertension), and 82.28% (health/nephropathy).
• Demonstrated superior multi-classification performance compared to state-of-the-art methods using support vector machine-radial basis function.
• Obtained a smaller representation error in distinguishing between healthy and diseased pulse signals.

Conclusions:

• The proposed Gabor function-based decomposition method offers superior performance in pulse signal analysis.
• This method effectively distinguishes between healthy and diseased individuals based on pulse wave characteristics.
• The approach shows significant potential for computerized disease diagnosis and health status monitoring.