Medical Image Classification By Genetic Algorithm And Ensemble Tree Learning Model

Skin cancer represents a major global health concern, impacting millions of people across the world. This makes timely detection and precise diagnosis extremely important, with dermoscopic imaging serving as an effective tool for identifying abnormalities at an early stage. In this study, a novel classification model is presented to distinguish skin medical images as either normal or abnormal. The proposed framework is divided into two main modules. The first module focuses on enhancing image quality through noise elimination and identifying the most affected regions that contribute significantly to diagnosis. The second module is responsible for extracting histogram-based and CCM features from the processed images, which are then utilized to train the Ensemble Tree classifier. The experimental evaluation was carried out using a real-world skin cancer image dataset. The obtained results demonstrate that the MICAIML improves detection accuracy with existing approaches.