Features Extraction Technique for ECG Recording Paper

Generally the ECG is recorded on a thermal paper which cannot be stored for a long time, because thermal trace over time becomes erased gradually. However some hospitals are saving the ECG thermal papers as scanning images in the electronic equipments (like computers) to maintain medical records, but this method needs to high memory capacity, and use less scanning resolution that gives signal accuracy is less at preview. In this paper image processing techniques are developed for an electrocardiogram (ECG) feature extraction and signal regeneration as a digital time series signal. The 12-lead ECG signals extracted from the recording paper and converting it to a digital time series signals. Feature extraction and the digital time series signal were tested on 30 of 12-lead ECG paper records from the MIT-BIH arrhythmia database, and the accuracy was between 96.31% and 98.25%. In addition this techniques also can be used for features extraction to perform an automatic heart disease classification using one of the artificial intelligence methods.