Reduction of motion artifacts from Pulse oximeter's PPG signals using MSICA

Pulse oximeter employed in intensive care units and is a crucial equipment employed to measure the vital parameters like heart rate and blood oxygen saturation levels. Using Pulse oximeter, photoplethysmographic (PPG) data is recorded using the PPG sensors attached to the patient is used to estimate oxygen saturation levels. Patient movements while recording the PPG data may result in erroneous estimation of required parameters may result in wrong diagnosis by the clinician. For exact estimation of arterial blood oxygen saturation (SpO2), a clean motion artifacts (MA) reduced is required. MA can be detached from the recorded PPG data using a classical filters, but the in-band noise components cannot be eliminated. In this work, we present a simple technique, multi-scale independent component analysis (MsICA), to reduce MA component present in PPG data. The authors proposed the technique by considering its advantages of MSICA when applied to quasi stationary signals. The superiority of proposed method has been proved by comparing the experimental results with results obtained using Independent component analysis (ICA) method. PPG data recorded with different MA (vertical, horizontal and bending movements of patient's finger) is considered for experiment analysis. Obtained SpO2 parameter computed values proved the effectiveness of the technique in measurement of accurate SpO2, helpful for effective medical diagnosis.