Tiny ML-Based Non-Invasive Approach of Cardiac Monitoring

Cardiac monitoring is crucial for health assessment, and in the era of technology, the integration of cutting-edge technologies has opened up new opportunities for non-invasive and remote monitoring solutions. The focus of this study is cardiac monitoring use-case involving Tiny Machine Learning (TinyML). Sensor data from Electrocardiogram(ECG) and Photoplethysmography(PPG) waveforms, which are collected and pre-processed such as signal cleaning and robust feature extraction to ensure that the acquired data are kept intact also we introduce a 1-D Convolutional Neural Network for efficient estimation of heart modalities. We introduced comprehensive analysis that includes time, frequency, and nonlinear domains which proves to be a critical method to validate our data sets and aim to study the underlying HRV metrics. This task of remote monitoring and deployment is handled by the Arduino Nano 33 BLE Sense, which uses Tiny Machine Learning's capabilities for real-time heart state classification. To summarize the applications of wearable ECG & PPG devices, specifically their analysis, identify their potential use case for remote applications using Edge AI, and outline our future research directions to realize their full potential to prevent and diagnose CVDs.