A Novel and Efficient Framework for Diagnosing ECG Signals Based on the Digital Signal Processing and Optimized Transformer Model

• DOI: 10.17725/j.rensit.2024.16.239
• Publication Date: 2024-04-24T19:52:02Z
• Authors: Anas F. Ahmed; Khalida S. Rijab; ; Ahmed T. Kamil

Heartbeat disorders are considered one of the main maladies that cause mortality. Therefore, their precocious diagnosis via ECG signal is critical for introducing prompt therapy. The advanced automatic classification signals has potential to save cardiologists a tremendous amount time while simultaneously decreasing chance misdiagnosis. dilemma massive parameters troubling current methods classification. Most recent exhibit inadequate performance diagnosing in inter-patient mode. In an attempt deal with above limitations, this study offers innovative, efficient, and end-to-end model. suggested model uses optimized transformer framework classify heartbeats according "Association Advancement Medical Instrumentation, AAMI," obeys setting. We constructed efficient architecture called network substitute Self Attention Unit (SAU) encoder part model, which includes network, outperforms SAU-based requires fewer computations. A robust embedding based on Convolutional Neural Network (CNN) Squeeze Excitation (SE) network-based attention scheme been used weighting Local Shape Pattern (LHSP) features presented. introduced exceeds state-of-the-art. An extensive test done compare achievements those cardiologists. results proved closeness performances.