Intelligent Estimation of Non-Invasive Blood Pressure From the Electrocardiogram Signal (ECG)

Continuous real-time blood pressure monitoring offers remarkable potential for the detection of cardiovascular disease. In fact, detection from an inflatable cuff makes the patient uncomfortable as a consequence of the inflation pressure of the cuff on the arm, which causes failures in detecting the results or errors in interpreting the systolic blood pressure values and diastolic blood pressure. In this contribution, we present an automatic system to predict the blood pressure signal based on the complexity analysis of the ECG signal, where the features extracted are signal mobility, signal complexity, fractal dimension, entropy, and autocorrelation with the neural network approach (ANN). Our technique is mainly tested on the MIMIC-III databases available to the public on physionet to have the ECG and ABP signals for feature extraction. The results provided argue in favor of our method in terms of efficiency and precision. Indeed, the proposed approach offers the possibility of detecting arterial pressure such as systolic arterial pressure and diastolic arterial pressure in real time