Evaluation of Spectral Trimming Technique for its Employment in Clinical Systems

Electrocardiogram (ECG) is a standard diagnostic tool for the diagnosis of various heart conditions and evaluating the effectiveness of heart treatments. Similarly, primary purpose of stress ECG is to detect ischemia, which may occur during physical exertion. In both the methods, significant noise such as powerline interference and motion artifacts get added which pose difficulty while analysing ECG signal. There are a few well known techniques such as Adaptive Noise Cancellation, Adaptive Line Enhancement, Wavelet denoising etc for suppressing these artifacts. The optimum performance of these methods depends on the selection of several input parameters. To overcome the subjective behaviour of these methods, Spectral Trimming Technique is introduced for suppressing the artifacts without compromising morphological features significantly. We worked further to evaluate the diagnostic features extracted to prove the efficacy of this method in clinical environment. In this paper we have presented the quantitative analysis of this evaluation.