1D CNN Framework on ECG Signals

This study delves into the realm of Biomedical Signal Processing, a field where Convolutional Neural Networks (CNNs) are renowned for their adeptness at amalgamating feature extraction and classification. However, there is a main challenge associated with the use of 2D CNNs, which is their computational complexity. This computational burden can render them impractical for real-time applications and the analysis of 1D signals. To overcome this challenge, this project employs 1D CNNs, which exhibit a more efficient computational profile and heightened performance, making them particularly well-suited for a broad spectrum of 1D signal processing tasks. The focal point of this endeavor is the application of the 1D CNN framework in the domain of Biomedical Signal Processing, with a specific emphasis on Electrocardiogram (ECG) signals. The principal objective is to conduct a comparative assessment of the performance between two 1D CNN models on different sizes datasets. This comparison will serve to illuminate the potential of 1D CNNs in optimizing the analysis and classification of 1D biomedical signals, with a particular focus on ECG signals. In a broader context, this research endeavors to make a substantial contribution to the enhancement of the efficiency and accuracy of biomedical signal processing techniques. Its significance is especially pronounced in the context of heart health monitoring, wherein ECG signals play a pivotal role. By harnessing the power of 1D CNNs, this project aims to pave the way for more effective and precise biomedical signal processing, ultimately contributing to advancements in the field of heart health monitoring through ECG signal analysis. Our 1D CNN model was able to achieve 98.6% accuracy using ReLu as the activation function and 98.65% accuracy using GeLu as the activation function.