ArrythML: An Autoencoder-Based TinyML Approach for On-Device Arrhythmia Detection on Resource-Constrained Embedded Systems
2026-06-01 • Machine Learning
Machine Learning
AI summaryⓘ
The authors created tiny machine learning models that can analyze heart signals (ECG) and detect irregular heartbeats right on small devices with limited power and memory. They tested their models using a well-known heart rhythm dataset and made sure the models work both on computers and on a tiny microcontroller. Their best model correctly finds most irregular beats quickly and fits into very little memory, showing it can run on wearable gadgets without needing internet. They also found that some mistakes by the model might actually highlight early signs of issues not marked in the original data. This work shows it's possible to do real-time heart monitoring privately on small, low-power devices.
ECG segmentationarrhythmia detectionTinyMLautoencoderINT8 quantizationMIT-BIH Arrhythmia DatabaseESP32-S3TensorFlow Lite Microembedded systemsinference latency
Authors
Nagarajan S, Kurian Polachan
Abstract
Our work presents a method for ECG segmentation and arrhythmia detection using Tiny Machine Learning (TinyML) models for real-time, on-device inference on resource-constrained embedded systems. We develop INT8 quantized autoencoder-based TinyML models with minimal layers and parameters for embedded deployment. These models are evaluated using a custom dataset derived from the MIT-BIH Arrhythmia Database and validated in both PC-based simulations and on-device environments. For the evaluations, over 95,000 ECG segments are processed on an ESP32-S3 microcontroller running the TensorFlow Lite Micro runtime. Post-evaluation, detailed analysis, including annotation-wise and record-wise failure analysis, is conducted to characterize model behavior across diverse ECG morphologies and rhythm patterns and to explain missed detections. In several cases, apparent misclassifications may correspond to early or subtle anomaly patterns labeled as normal in the reference annotations, highlighting the model's sensitivity. A refined evaluation by filtering out ambiguous cases in the dataset shows that the best-performing DNN-based autoencoder achieves a recall of 84%, an F1-score of 79%, a model size of approximately 180 KB, and an inference latency of 9 ms on-device. These results demonstrate the feasibility of low-power, privacy-preserving embedded wearable systems capable of performing accurate arrhythmia detection entirely on-device.