Filtering Parameters Selection Method and Peaks Extraction for ECG and PPG Signals

Abstract

With the growth of electronic E-health, wearable devices have highlighted due to its practicality and comfort in sensing of the personal data. Those devices gather biosignals for heart rate measurements, blood oxygenation checks, and identification. From sensors in the devices, electrocardiogram (ECG), photoplethysmogram (PPG) generate unique identifiers for use in identifying people, just like fingerprint, faceid, iris, among others. Every process with sensors has noises, electromagnetic waves, and movement that can interfere with the system when identifying, analyzing, and checking the individual. From this, filtering is an indispensable step in any process. The identification is divided into stages, and the main and essential is the filtering because where preprocessing starts. Thus, this work proposes a method of select filtering parameters for ECG and PPG signals and peaks extraction to the purpose the apply them for any dataset, following a sequence of steps, filtering, characteristics extraction (Peaks) and to affirm our model the correlation between filtered and raw waves performing a wave overlap. It is achieving an 80% correlation between raw waves and filtered waves.