ArrythML: An Autoencoder-Based TinyML Approach for On-Device Arrhythmia Detection on Resource-Constrained Embedded Systems

2026-06-01

AI Summary

This study addresses the demand for real-time, low-power, and privacy-preserving arrhythmia detection on embedded devices by proposing an extremely lightweight INT8-quantized autoencoder model deployed directly on an ESP32-S3 microcontroller using TensorFlow Lite Micro. Leveraging a custom dataset derived from MIT-BIH, the approach achieves a recall of 84% and an F1 score of 79% with a model size of only 180 KB and an inference latency of 9 ms. The work substantially reduces computational overhead while demonstrating that certain instances labeled as false positives actually correspond to early or subtle cardiac anomalies, thereby highlighting the model’s ability to strike a favorable balance between sensitivity and practicality in resource-constrained settings.

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.

Problem

• arrhythmia detection
• TinyML
• on-device inference
• resource-constrained embedded systems
• ECG segmentation

Innovation

• TinyML
• Autoencoder
• On-device inference
• ECG arrhythmia detection
• INT8 quantization