Extracting GLCM Matrix from Images and Computing Haralick's 14 Feature Values

In this article, we need to extract the Gray-Level Co-occurrence Matrix (GLCM) from images and compute the 14 feature values proposed by Haralick. These feature values include: 1. Contrast: Measures the intensity contrast between a pixel and its neighbors, implemented by calculating the weighted sum of squared differences between pixel pairs in the GLCM. 2. Correlation: Evaluates the linear dependency of gray levels between pixel pairs, computed using statistical covariance and standard deviation measurements from the GLCM. 3. Energy (Angular Second Moment): Represents the uniformity of gray-level distribution, calculated as the sum of squared GLCM elements. 4. Inverse Difference Moment (Homogeneity): Measures the local homogeneity of texture, with higher weights given to closer gray-level pairs in the GLCM calculation. 5. Entropy: Quantifies the randomness or uncertainty in the image's gray-level distribution, computed using the probability distribution from the GLCM. 6. Autocorrelation: Assesses the self-similarity of pixel values at specific spatial relationships, derived from the GLCM pattern analysis. 7. Cluster Shade: Measures the skewness of the GLCM distribution, indicating the presence of asymmetric patterns in pixel clustering. 8. Cluster Prominence: Evaluates the peakedness of the GLCM distribution, reflecting the symmetry of pixel clustering patterns. 9. Maximum Correlation Coefficient: Determines the highest linear relationship between pixel pairs, calculated through secondary statistical analysis of the GLCM. 10. Contrast Entropy: Combines contrast and entropy measures to evaluate texture variation complexity in the gray-level distribution. 11. Information Measure of Correlation 1: Assesses non-linear correlations between pixels using information-theoretic approaches based on joint entropy calculations. 12. Information Measure of Correlation 2: Provides an alternative measure of non-linear pixel dependencies through entropy-based correlation analysis. 13. Maximum Probability: Identifies the most frequent gray-level pair combination in the image, corresponding to the highest value in the GLCM. 14. Sum Average: Computes the average of pixel cluster intensities, derived from the marginal distribution sums of the GLCM. These 14 Haralick features can be obtained by computing the GLCM matrix from images and are widely used for texture analysis, pattern recognition, and image processing applications. In implementation, these features are typically calculated using specialized image processing libraries that provide GLCM computation functions with configurable parameters for distance and angle relationships between pixels.