Feature Extraction Using Gray-Level Co-occurrence Matrix and k-Nearest Neighbor Classification

In this study, we employed Gray-Level Co-occurrence Matrix (GLCM) for texture feature extraction from images, followed by classification using the k-Nearest Neighbor algorithm. Our implementation includes specific functions: creat_apprentissage prepares training samples by organizing image data and corresponding labels, cooccurence computes GLCM features by analyzing pixel intensity relationships at specified distances and orientations, knn implements the classification algorithm using distance metrics to find closest training samples, and classif performs the final texture image categorization. This approach yields more accurate results in image classification by effectively capturing texture characteristics. The GLCM implementation typically calculates statistical features like contrast, correlation, energy, and homogeneity from the co-occurrence matrix. The k-NN algorithm classifies images based on feature similarity with training data. Additionally, this methodology enables the study of similarity between different textures and facilitates understanding of their distinctions. Future research could explore more texture-based image classification techniques applicable to various domains such as medical imaging, geological image analysis, and remote sensing. This will enhance our ability to interpret and utilize image information while developing more accurate and reliable classification algorithms.