Hybrid Features of Mask Generated with Gabor Filter for Texture Analysis and Sobel Operator for Image Regions Segmentation Using K-Means Technique

Abstract

To make the image easily represented for more analysis and processing the segmentation procedure is required, where the image is portioned into its formed regions using some segmentation techniques based on features extraction. In this paper, a proposed procedure for finding the regions that formed the image is achieved based on hybrid features in two different components of different two colors spaces L*a*b* and RGB segmented by the k-means method. The hybrid features which comprise the mask segmentation are a combination of texture image characterization extracted by the Gabor filter and gradient image intensity by the Sobel operator after image quality enhancement by applying wiener filter noise reduction and contrast enhancement using Contrast limited adaptive equalization (CLAHE). Some statistical metrics are used for evaluating the performance of the proposed work stages.