Images Segmentation Method on Comparison of Feature Extraction Techniques

An algorithm of segmentation by using feature extraction techniques of Synthetic Aperture Radar (SAR) images in this paper. The segmentation processor are shown to be of interest for analysing SAR image data. The extracted and selected features are then used to train different neural-network based classifiers. Segmentation makes use of wavelet decomposition and unsupervised clustering based on PCA. The learning approach of neural networks is used for combining various features of different areas of an image. The outcomes of the proposed segmentation techniques are compared to the standard Gaussian discriminant analysis in the case of a real E-SAR image.