ECG VisionSense: A Cloud-Deployed, Flask-Based Intelligent Cardiovascular Disease Diagnosis System

The Echocardiogram (ECG) classification has great importance due to prevailing nature of ailments related to heart. The remote monitoring devices which record the heart activity are interpreted by physicians. This paper focuses on the need to provide a reliable method for automatic ECG classification. By employing convolution neural network, the model is trained on the dataset generated using PTB-XL data. The ability to enhance and recognize the intricate patterns of the signals in the generated dataset, makes it capable of improving the ECG classification. The standalone ECG signal images that contain 12 leads have been classified according to their patterns. The deep learning network showed the results with accuracy of over 99.04% for the training and 96.10% for validation on the generated data. The model is deployed on Amazon Web Services (AWS) cloud platform which enables it to be accessible to the users. The deep learning technique along with the deployment on the cloud can help in assisting the physician to classify the ECG signals.