Please use this identifier to cite or link to this item:
|Title:||Acquisition, Processing, Coding & Study of Fractal Behaviour of ECG Signals|
|Abstract:||The electrocardiogram (ECG) is a method of recording the electrical activity of the heart. ECG signal is non-stationary signal including valuable clinical information, but frequently this information is corrupted by noise. So the signal needs to be noise free. Here, attempts have been made to filter the corrupted ECG signals with the help of filters and transformational techniques. Mainly two types of filters are discussed here: Butterworth band pass filter and Savitzky-Golay filter. The Savitzky-Golay filter proves to be efficient in filtering the noise. The transformational techniques that are used here are the Walsh transform and the Discrete Cosine transform. The transformed signal is filtered with the Savitzky-Golay filter. As the performance indices, here signal-to-noise ratio (SNR), signal-to-inference ratio (SIR) and percent-root-mean-square difference (PRD) are taken into consideration. It is obvious from the results that the Discrete Cosine transformation (DCT) is better in eliminating noise from the corrupted signal. The data of ECG signal is significantly very large. So in order to reduce the data, coding techniques are used here to reduce the ECG signal. The coding techniques mainly used here are Huffman Coding and Run Length Coding. At first , DCT is done upon the ECG signal and then the coding techniques are applied on the transformed signals. It is seen from the results that the, Run Length coding gives better results. Also the size of the ECG data has been reduced after encoding the signal. . After applying Huffman encoding and decoding, the morphology of the ECG signal is lost, but on the contrary, the Run Length encoding and decoding gives better results. In this case the morphology of the signal is not lost. A concept of fractal is also applied on the ECG signal. Fractal is a mathematical investigation for characterizing complex replicating geometric patterns at different scale lengths. Here fractal behavior of the ECG signal, both normal and diseased, has been studied using the Rescaled Range Analysis and Multifractal Detrended Fluctuation Analysis (MF-DFA). The fractal dimension has been found using the Rescaled Range Analysis. All the results are implemented on the MATLAB platform.|
|Appears in Collections:||Bioscience & Engineering - Master's Degree Theses|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.