Method and apparatus for hypertension classification

A method for real-time evaluation and screening of hypertension of a subject to be performed in a system comprising a processor and storage storing data and software, comprising steps of;

• obtaining an original infrared photoplethysmograph (PPG) signal of the subject using a fingertip PPG;
• obtaining a plurality of derivative waveforms as a function of the PPG signal, including at least a velocity of plethysmograph (VPG) waveform and an acceleration of plethysmograph (APG) waveform;
• defining and extracting a plurality of waveforms from the PPG signal and the derivative waveforms of the PPG signal;
• analyzing a physiological significance of the extracted plurality of waveforms to obtain at least one physiologically significant feature;
• selecting the at least one physiologically significant feature from the extracted plurality of waveforms, the at least one physiologically significant feature selected from a group including;
• a slope from a first point of the PPG signal corresponding in time to a minimum of the APG waveform immediately preceding a systolic maximum of the PPG signal and a second point of the PPG signal corresponding in time to a maximum of the APG waveform immediately following the systolic maximum of the PPG signal;
• or
• a time interval from the systolic maximum of the PPG signal to the second point of the PPG signal;
• and
• using a cubic Support Vector Machine (SVM) classification model of hypertension to generate classification of patients into different stages of hypertension.

A system for real-time evaluation and screening of hypertension of a subject, comprising:

• a processor and storage storing data and software for performing a method comprising steps of:
• obtaining an original infrared photoplethysmograph (PPG) signal of the subject using a fingertip PPG;
• obtaining a processed PPG signal by performing a signal conditioning of the original infrared PPG signal;
• obtaining derivative waveforms of the processed PPG signal including a velocity of plethysmograph (VPG) and an acceleration of plethysmograph (APG) from the processed PPG signal;
• defining and extracting waveforms from the processed PPG signal and the derivative waveforms of the processed PPG signal, the extracted waveforms comprising at least a time interval, an amplitude, an area, a ratio, a slope, and an energy parameter;
• analyzing a physiological significance of the extracted waveforms to obtain at least one physiologically significant feature;
• selecting the at least one physiologically significant feature from the extracted waveforms, the at least one physiologically significant feature selected from a group including:
• a slope from a first point of the processed PPG signal corresponding in time to a minimum of the APG waveform immediately preceding a systolic maximum of the processed PPG signal and a second point of the processed PPG signal corresponding in time to a maximum of the APG waveform immediately following the systolic maximum of the processed PPG signal;
• or
• a time interval from the systolic maximum of the processed PPG signal to the second point of the processed PPG signal;
• using a cubic Support Vector Machine (SVM) classification model of hypertension to generate classification of patients;
• and
• displaying the generated classification of patients into different stages of hypertension in the display of a portable device.