M.Sc. Tezi Görüntüleme

      Every natural event is often repetitive. Fourier transform is used to find global frequency of a signal which characterizes the corresponding natural event. However, the Fourier transform is not useful for determining the frequency of a signal at a given time since it considers the signal as global. In order to find the frequency at a given time interval, several variants of the Fourier transformation are employed. One of them is Short-Time Fourier transform. The Short-Time Fourier transform performs Fourier transform on the signal broken up into predefined windows of size well selected. Narrower or wider windows give rise to resolution problems. One of the methods which determines the frequency at a given time t is the Wavelet transform. The Wavelet transform is introduced to overcome the resolution problem in Short-Time Fourier transform. It has advantages to easily decompose a signal into its components (low and high frequency). A better texture segmentation can be achieved using frequency information of the image as well as other image information. Therefore the Wavelet transform provides a facility to obtain time-frequency localization.

      The resonance algorithm is used for classification of extracted texture feature vectors. Beside its segmentation ability, the resonance algorithm is also capable of wiping out the effects of images which contain wide-ranged graduations in intensity due to environmental illuminations.

      Key words: Texture Segmentation, Wavelet Transform, Fourier Transform,