Quanvolution Neural Network to Recognize arrhythmia from 2D scaleogram features of ECG signals

Quantum computing is the main emerging technology solving complex problems and even though error raised will be high, can be computed using customized algorithms. We propose the use of discrete wavelet transform to decompose the ECG signals followed by computing 2D scalogram to obtain time-frequency features and apply Quanvolutional Neural Network to classify those scalogram images to recognize Arrhythmia. This is the first paper to introduce scalogram and Quanvolutional neural networks. We considered using publicly available physio net MIT-BIH arrhythmia database for our research. The proposed model of hybrid quantum classical model comprising quantum convolutional neural networks for the MIT-BIH arrhythmia database resulting in the precision of 98% and Receiver Operating Curve Score of 100%.