ارائه روشی بر مبنای تبدیل دو خطی سودو ویگنر ـ ویل جهت تشخیص آریتمی های قلبی

This study focuses on the automatic identification of idioventricular rhythm (IVR), atrial flutter (AFL), and atrial fibrillation (AFIB) through advanced electrocardiogram (ECG) signal processing using the MIT-BIH database. Initially, ECG signals were preprocessed to eliminate power line interference, muscle motion artifacts, and baseline wander, while preserving the intrinsic morphology of the cardiac waveform. Following denoising, R-peaks were accurately detected using an optimized algorithm and served as temporal anchors for subsequent analysis. Based on these reference points, Q and S waves were extracted, and the precise QRS complex boundaries were determined. The onset and offset of the Q and S waves were used to define the ventricular activity interval, while a similar procedure was applied to the P wave to separately analyze atrial activity. The innovation of this study lies in its hybrid signal decomposition framework, which combines precise time-domain segmentation with frequency-domain feature extraction. Unlike conventional methods that separate components solely in the frequency or time–frequency domain, the proposed approach first isolates ECG components (P, QRS, and T waves) in the time domain based on accurate morphological boundaries, ensuring non-overlapping frequency characteristics. Subsequently, frequency-based features are extracted from each individual component, significantly enhancing discriminability and classification accuracy. This dual-stage design effectively reduces waveform interference and outperforms traditional time–frequency analysis techniques. After component separation, a representative beat containing P and QRS waves—typically the fourth or fifth beat—was selected, and the pseudo-Wigner–Ville (PWV) distribution was applied to achieve a high-resolution time–frequency representation. Statistical and time–frequency features derived from PWV results and wave boundaries were employed as discriminative indicators for rhythm classification. The results demonstrate that the proposed framework provides a novel, efficient, and accurate methodology for automatic cardiac arrhythmia detection, offering strong potential for intelligent cardiac monitoring and screening systems.