What is the clinical evidence from this study?

Study design: Other. Population: Arrhythmia. Intervention: ArythmiAR system (Deep learning ECG classification with AR 3D visualization). Primary outcome: ECG classification accuracy.

What question did this study set out to answer?

The aim is to enhance the accuracy and timeliness of arrhythmia diagnosis using ECG and augmented reality technology.

May 16, 2026Open Access

Enhancing Arrhythmia Diagnosis Through Ecg Deep Learning Classification Deploying and Augmented Reality 3d Heart Visualization and Interaction

Key Result

The ArythmiAR system, utilizing a Multilayer Perceptron model, achieved 99.07% accuracy in classifying ECG signals for arrhythmia detection on the PhysioNet MIT-BIH dataset.

Key Points

The aim is to enhance the accuracy and timeliness of arrhythmia diagnosis using ECG and augmented reality technology.
Developed a system called ArythmiAR integrating CNN and AR for ECG classification and 3D visualization.
Applied data rebalancing techniques to improve model performance based on the PhysioNet MIT-BIH Arrhythmia dataset.
Focused on MLP and CNN models for effective classification of ECG data.
Achieved an accuracy of 99.07% with the MLP model on the PhysioNet MIT-BIH Arrhythmia dataset.
Demonstrated the effectiveness of AR for visualizing heart sub-regions related to arrhythmias.
Provided clinicians with interactive tools for better diagnosis and treatment planning.

Structured PICO

Population

PhysioNet MIT-BIH Arrhythmia dataset

Intervention

ArythmiAR system combining Convolutional Neural Networks (CNN), Multilayer Perceptron (MLP), and Augmented Reality (AR) for ECG classification and 3D heart visualization

Outcome

Accuracy of ECG classification

A novel system combining deep learning for ECG classification and augmented reality for 3D heart visualization achieved 99.07% accuracy in arrhythmia detection on a standard dataset.

Limitations

Lack of real-time monitoring
Lack of multi-lead ECG analysis

Abstract

Cardiovascular diseases (CVDs) are the leading cause of death worldwide, which emphasises the importance of timely and accurate diagnosis. Electrocardiography (ECG) is a crucial diagnostic tool that scrutinises the electrical activity of the heart to identify and monitor diverse cardiac conditions. However, manual detection of ECG features and categorisation of heartbeats is a difficult and time consuming task which requires a lot of skill. To overcome this problem, we have developed a new system called ArythmiAR, which combines Convolutional Neural Networks (CNN) and Augmented Reality (AR) for interactive diagnosis with 3D visualisation and interaction in real time. ArythmiAR has a number of new key features including: deep learning based ECG classification for accurate detection of arrhythmia, 3D heart modelling and assembly for better visualisation, an augmented reality (AR) interface to deploy CNN models, 3D localisation of heart sub-regions responsible for arrhythmia anomalies and better 3D visualisation and interaction. The paper discusses various methods for ECG classification by applying data rebalancing techniques to improve the models' performance. The study focuses on Multilayer Perceptron (MLP) and Convolutional Neural Network (CNN) models, which yielded highly competitive results on the PhysioNet MIT-BIH Arrhythmia dataset, achieving an accuracy of 99.07% with the MLP model. In this work we also consider the use of the ECG classification deep learning model in an augmented reality setting, a prototype of augmented rendering which allows the user to locate, visualise and interact with specific parts of the heart responsible for arrhythmias. This platform gives doctors the tools to make better diagnoses and better treatment plans, improving the care of all patients

Bookmark

View Full Paper