Ischemia Detection: By Identification of Isoelectric Line and ST Segment

Abstract

In this research work we present an algorithm for the automated detection of ST deviation that can be useful in diagnosing Coronary Heart Disease (CHD) using electrocardiogram (ECG) recordings. The technique is developed using Long Term ST database (LTST DB). Preprocessing is carried out prior to the extraction of ST segment which includes noise filtering using seven-point parabolic filter and then application of Wavelet Transform for QRS detection. The algorithm determines the R-peak detection in large number of samples, and then estimates the ST-segment’s relative level with respect to iso-electric level using overlapping band selection method. It then compares the two levels, which is later used for ischemia detection. The performance of the proposed solution was evaluated on 14 records from LTST database. We found that there is reasonable amount of accuracy (98.5%) and it can therefore be concluded that the algorithm which have proposed can be used for most of the practical purposes.