Fragmented QRS Rapid Identification and Feature Extraction for Myocardial Infarction Diagnosis

Early and accurate diagnosis of myocardial infarction is crucial for improving prognosis, yet traditional electrocardiogram indicators (ST-segment elevation, pathological Q waves) have shown declining sensitivity with the widespread adoption of interventional therapies. This study proposes an intelligent framework integrating wavelet-based multiscale dynamic analysis and extremum-point morphological rules to achieve automated rapid identification and feature extraction of fragmented QRS (fQRS) complexes. The algorithm employs multistage preprocessing to suppress noise interference, enabling precise QRS waveform localization and signal enhancement. A wavelet analysis-based fQRS recognition algorithm is developed for rapid identification and screening of fragmented QRS patterns in ECG signals. For identified ECG signals containing fQRS waveforms, we further design dual-threshold extremum point discrimination rules to encode clinical expertise into quantifiable morphological features, facilitating fragmented QRS waveform extraction across all leads. Validation is performed using the Physionet PTB database. This research overcomes the subjectivity bottleneck of manual interpretation and provides a highrobustness solution for non-invasive diagnosis of myocardial infarction and personalized prognostic stratification.