Angelia Ayu Chandra, Cecilia Sunnia, Kenrick Alvaro Wijaya, Abdi Dharma, Arjon Turnip, Mardi Turnip

ECG-BASED ARRHYTHMIA DETECTION USING THE NARROW NEURAL NETWORK CLASSIFIER

Introduction

Ecg-based arrhythmia detection using the narrow neural network classifier. Detect arrhythmias efficiently on wearable devices using the lightweight Narrow Neural Network Classifier. This ECG-based model achieves 98.9% accuracy for real-time detection.

Abstract

Electrocardiograms (ECG) are important for detecting arrhythmias. Conventional models such as CNN and LSTM are accurate but require large amounts of computation, making them difficult to use on wearable devices and for real-time monitoring. This study evaluates the Narrow Neural Network Classifier (NNNC) as a lightweight and efficient alternative. The dataset consists of 21 subjects with 881 ECG samples, categorized based on walking, sitting, and running activities, and processed through bandpass filtering, normalization, and P-QRS- T wave segmentation. The data is divided into training (70%), validation (15%), and test (15%) sets. The NNNC has 11 convolutional layers, a ReLU activation function, a Softmax output, and 120,000 parameters. The model was trained using the Adam optimizer, a batch size of 32, and a learning rate of 0.001 for 100 epochs and compared with SVM, CNN, and LSTM using accuracy, precision, recall, F1-score, and ROC-AUC. The results show that NNNC achieves an accuracy of 98.9%, a precision of 99.2%, a recall of 99.2%, and an F1-score of 99.2%, higher than SVM and comparable to CNN/LSTM, with lower computational consumption. The model is capable of reliably detecting early arrhythmias. These findings support the potential of NNNC for ECG-based automatic diagnostic systems, including real-time implementation on wearable devices, although further research is needed for large-scale validation

Review

The manuscript "ECG-BASED ARRHYTHMIA DETECTION USING THE NARROW NEURAL NETWORK CLASSIFIER" addresses a critical challenge in medical diagnostics: enabling accurate and efficient arrhythmia detection on resource-constrained devices like wearables for real-time monitoring. The authors propose the Narrow Neural Network Classifier (NNNC) as a lightweight alternative to more computationally intensive conventional models such as CNNs and LSTMs. This focus on efficiency without compromising accuracy is highly relevant and timely given the increasing demand for personalized, continuous health monitoring solutions. The initial findings presented in the abstract suggest a promising approach to overcome existing barriers to widespread adoption of AI-powered ECG analysis. The methodology outlines a standard machine learning pipeline, including bandpass filtering, normalization, and P-QRS-T wave segmentation on a dataset of 881 ECG samples from 21 subjects, categorized by activity. The NNNC architecture comprises 11 convolutional layers with 120,000 parameters, trained using an Adam optimizer. Crucially, the NNNC's performance, as measured by accuracy, precision, recall, and F1-score (all around 99%), is reported to be superior to SVM and comparable to CNN/LSTM, with the significant advantage of lower computational consumption. While these metrics are impressive, the abstract's brevity leaves some critical details open for further scrutiny. Specifically, the relatively small dataset size (21 subjects, 881 samples) warrants careful consideration regarding the generalizability of the results. Furthermore, the abstract mentions "early arrhythmias" and "categorized based on walking, sitting, and running activities," which needs clarification regarding the specific types of arrhythmias detected and whether the dataset includes pathological conditions or primarily healthy individuals performing activities. Quantification of the "lower computational consumption" would also strengthen the claims. The potential implications of this work are substantial, particularly for advancing automated diagnostic systems and facilitating real-time arrhythmia detection on wearable devices, which could lead to earlier intervention and improved patient outcomes. The claim of reliably detecting "early arrhythmias" is particularly impactful, if adequately substantiated. However, as the authors rightly acknowledge, large-scale validation is essential to confirm the robustness and generalizability of the NNNC across diverse patient populations and arrhythmia types. For the full paper, a detailed discussion on the specifics of the arrhythmia classes, the dataset's clinical relevance, and a quantitative comparison of computational resources (e.g., inference time, memory footprint) against CNN/LSTM would significantly enhance the contribution. Overall, the abstract presents a compelling case for a novel and efficient approach to ECG-based arrhythmia detection, warranting further exploration and detailed exposition in a full manuscript.

Full Text

Comments

You need to be logged in to post a comment.