2. ABOUT GLCM
Feature extraction can be done using many different techniques, of which
Gray level co-occurrence matrix (GLCM) is the widely used method. GLCM
has the second order statistical data of neighboring pixels of an image.
The details of the image content can be obtained from GLCM through
textural properties. It also shares the information regarding the
association of the neighboring pixels to each other. The difference in
patterns of image pixel values in an image is given accurately by GLCM.
Patterns are certain kind of statistics that are used on each GLCM.
(i) Gray level occurrence matrix
These matrices are one which gives information about the presence of a
texture in an image. The information about their presence that is
uniform or irregular presence of the texture is well explained in this
matrix. These matrices are invariant to rotation.
(ii) Haralick Statistics
The design of Haralick texture classification is that this statistics
gives information regarding the general design of the pattern of
neighboring pixels in a given image.
(iii) Gabour Features
The Gabor wavelets are built using the Gabor filter method. In this
method, a band pass filter is applied to the input at different
orientations and scales, similar to the mechanism used by the human
visual system.
The basic steps involved in this process are,
1. Quantization
2. Creation of GLCM
3. Calculation of feature