شماره ركورد كنفرانس :
3540
عنوان مقاله :
Contourlet-Based Levelset SAR Image Segmentation
پديدآورندگان :
Alireza Ebrahiminia Department of Electrical and Electronic Engineering - Shiraz University of technology , Mohamad.S Helfroush Department of Electrical and Electronic Engineering - Shiraz University of technology , Habibollah Danyali Department of Electrical and Electronic Engineering - Shiraz University of technology , Shabab Bazrafkan Department of Electrical and Electronic Engineering - Shiraz University of technology
كليدواژه :
Contourlet , kurtosis , Segmentation , Levelset , SAR
عنوان كنفرانس :
همايش بين المللي هوش مصنوعي و پردازش سيگنال
چكيده لاتين :
Segmentation in Synthetic Aperture Radar (SAR) images has become very important in recent years because of its wide range of applications. There are lots of methods introduced for this purpose in literature. In this paper we used contourlet as a new method for feature extraction. Here the levelset meth-od is used with a statistical based kernel. The parameters of this kernel have been calculated using high order cumulants. For this purpose the estimation of energy distribution in Contourlet transform has been used. Our experimental re-sults show that the proposed method is effective in Image segmentation contain-ing inhomogeneity intensities such as SAR images and also outperforms wave-let based methods.
