A New Method for Texture Features Extraction (TFE) of Ultra Sound Image Based On Gabor Filter Segmentation Using Matlab

In this paper titled: “A New Method For Texture Features Extraction (TFE) Of Ultra Sound Images Based OnGabber Filter Segmentation Using Matlab” is used for to detect abnormal structural changes of ultra sound images. In research is being conducted with the objective to innovatively develop and apply image segmentation and image feature extraction techniques to efficiently segment the ultrasound image which can be used to automatically detect the diseases in different organs of the body. The process of identifying regions with similar texture and separating regions with different texture is used for the segmentation of ultrasound images.Firstly, a multi-channel texture analysis technique that relies on 2D Gabor Filters is used to isolate regions of perceptually homogeneous texture in an image. Textures are modeled as patterns dominated by a narrow band of spatial frequencies and orientation. Properly tuned Gabor filters react strongly to specific textures and weakly to all others. Then the features of the image were extracted from the image and the clustering of pixels in the feature space produced the segmented image. Unsupervised approach is used for texture segmentation. K-mean clustering method is proposed to cluster the pixels belonging to the same texture region provided the number of different textures in the image is known beforehand.