Identification of Dermatological Diseases using AI-Driven Entropy and Texture-Based Analysis

Dermatological diseases affect a significant portion of the global population. Traditional diagnostic methods such as visual inspection and biopsies are subjective, invasive, and time-consuming. To address these limitations, this study proposes an entropy-based texture analysis framework combined with Gray Level Co-occurrence Matrix (GLCM) features for the automated identification and classification of skin diseases using standard color dermatological images. The methodology involves pre-processing the input images through normalization and resizing, followed by the extraction of five key texture features: contrast, correlation, energy, homogeneity, and entropy. A comparative evaluation across four dermatological conditions Morgellons, Dermatitis, Psoriasis, and Vitiligo demonstrates that entropy and homogeneity are the most effective features in capturing disease-specific textures, whereas contrast, correlation, and energy exhibit limited discriminative capability. Furthermore, the study examines the impact of varying window sizes (5, 15, and 25) for texture extraction and identifies a 5×5 window as the optimal configuration for preserving critical lesion details. The proposed approach provides a lightweight, interpretable, and non-invasive solution that can serve as a valuable component in clinical decision-support systems. This work contributes to the advancement of AI-driven dermatological diagnostics by offering a cost-effective and accessible methodology for automated skin disease identification.