作者: Sandeep Raj , GSS Praveen Chand , Kailash Chandra Ray , None
DOI: 10.1016/J.MICPRO.2015.07.013
关键词:
摘要: This paper aims for accurate diagnosis of arrhythmia beats in real time to enhance the health care service cardiovascular diseases. The proposed methodology involves integration R-peak detection algorithm, FFT (fast fourier transform) based discrete wavelet transform feature extraction and feedforward Neural Network Architecture classify generic cardiac beat classes into eight categories namely Right Bundled Block, Left Preventricular Contraction (PVC), Atrial Premature (APC), Ventricular Flutter wave (VF), Paced Beat, Escape (VE) Normal beat. contributes development, prototyping analysis on ARM (Advanced RISC Machine) SoC (System-on-Chip) laboratory setup. system is validated by generating real-time ECG signals using MIT-BIH database while output monitored displaying device. performance implemented microcontroller computed performing experiment which achieves a high overall accuracy 97.4% with average sensitivity ( S e ) 97.57%, specificity p 99.59% positive predictivity P 97.93%. provides an assistive diagnostic solution users lead healthy lifestyle. Moreover, ARM-based can be fabricated handheld device reliable automatic monitoring condition heart patients.