Statistical Based Feature Extraction Texture Features What is

Statistical Based Feature Extraction

Texture Features What is texture? • Texture is a description of the spatial arrangement of color or intensities in an image or a selected region of an image.

Defining Texture There are three approaches to defining exactly what texture is: Structural: texture is a set of primitive texels in some regular or repeated relationship. Statistical : texture is a quantitative measure of the arrangement of intensities in a region. Modeling : texture modeling techniques involve constructing models to specify textures.

Grey Level Cooccurence Matrix • Co-occurrence matrix is a single level dependence matrix that contains the relative frequencies of two coordinate elements separated by a distance ‘d’. • The gray level co-occurrence can be specified in a matrix of relative frequency p (i, j) with which two neighboring resolution cells separated by distance d occur in the image, one with gray tone i and the other with gray tone j. • Generally, distances of one pixel and angles of 0º, 45º, 90º and 135º degrees are used. The (d=1, α=0º)-pixel pairs are horizontally adjacent, the (d=1, α=90°)-pairs are vertically adjacent.

GLCM Calculation Input Image P(i, j) (d=1, α=0º) P(i, j) (d=1, α=90º)

Features Description Entropy is a Statistical measure of randomness that can be used to characterize the texture of the input image Energy Provides the sum of squared elements in the GLCM. High values of Energy occur when the window is very orderly. Homogene It measures image homogeneity as it ity assumes larger values for smaller gray tone differences in pair elements. It is more sensitive to the presence of near diagonal elements in the GLCM. Contrast Returns a measure of the intensity contrast between a pixel and its neighbor over the whole. Measures the local variations in the graylevel co-occurrence matrix. The weights continues to increase exponentially as (ij) increases. Formula

Features Description Correlation Returns a measure of how correlated a pixel is to its neighbor over the whole image Range= [-1 1] Correlation is 1 /-1 for a perfectly positively or negatively correlated image. Correlation is Na. N for a constant image. Max. prob This simple statistic records in the centre pixel of the window. Max value occur if one combination of pixel dominates the pixel pairs in the window Mean The product of summation of the probability values in to their corresponding row or column Variance is a measure of the dispersion of the values around the mean of combinations of reference and neighbor pixels Formula

GLCM for the complete image Input image GLCM based texture image GLCM Computed for each pixel in an image Input image Entropy Inertia Homogeneity Autocorrelation Correlation Energy

Thank You