اختيار المويجة الأم المناسبة للتحويل المويجي : تطبيق على معالجة إشارة ECG

الملخص EN

One of the advanced transformations used in the digital signal processing is the wavelet transform, by which we can represent the digital signal into multiresolution levels.

Each level of the representation consists of two signals, the first is an approximation average signal and the second is a detailed signal, each approximation signal can be transformed in the next level of analysis into new two sub-signals: an approximation and detailed signals.

These multiresolution levels of details can be used for a wide variety of fundamental signal processing tasks, such as compression, removing noise or enhancing a recorded sound or image in various ways.

There are many types of wavelet basis (mother wavelets) can be used as a basis of wavelet transform.

The main purpose of this article is to answer the following question: how can we chooce the correct mother wavelet for wavelet transform of a specific kind of signals? We propose, in this article, a methodology of chooce the suitable mother wavelet for a specific kind of signals, we focus in the application on the ECG signals which are the most commonly used in this medical field.

The methodology presented in this article is applicable for wavelet transform to process and analyze any kind of non-random signals.

For this study, we used a set of standard ECG signals obtained from the international digital bio-signal databases ”www.physionet.org”.

This database is used by researchers to test their algorithms and develop their researches in the bio-signal field.

We used different kinds of normal and abnormal ECG signals with different duration and resolution.

We have tested 7 groups of the most famous mother wavelets used in both continuous and discrete wavelet transform.

These groups contain about 89 different type of mother wavelets.

The main aim of this study is to select the correct wavelet for the ECG signal analysis by wavelet transform.

Our results are evaluated using some qualitative and quantitative evaluation criteria used in the digital signal-processing domain, and we found that the bior3.1 is the most suitable mother wavelet for the used ECG Signals.