Deep Learning ECG Signal Analysis: Description and Preliminary Results

Electrocardiography is one of the most commonly performed examinations when a patient experiences symptoms of heart disease. A detailed analysis of the problem shows that it can also be used to detect and recognize emotions and activity. In this paper, we present several architectures for the analysis of electrocardiogram (ECG) data. Such architectures could be used for activity and emotion detection in augmentative and alternative communication (AAC) schemes. The use of deep learning algorithms shows the potential of using popular wearable devices and their usefulness for this analysis. Here, we present the preliminary results of our analysis, which provide promising conclusions on the application of a complete AAC framework. The presented methodology is applicable for the classification of the ECG data and can be further used with wearable devices. The conclusions and results of the study presented allow us to investigate real-time data analysis with a wearable device to predict certain cardiac conditions.