New Approach Generative AI Melanoma Data Fusion for Classification in Dermoscopic Images with Large Language Model

• Adriell Gomes Marques IFCE / LISIA
• Marcus Vinicius Candido de Figueiredo IFCE / LISIA
• José Jerovane da Costa Nascimento UFC / LISIA
• Cidcley Teixeira de Souza IFCE
• Carlos Mauricio Jaborandy de Mattos Dourado Junior UFC / LISIA
• Victor Hugo C. de Albuquerque UFC
• Luís Fabrício de Freitas Souzal UFCA / LISIA

Resumo

Skin cancer is a disease that causes thousands of deaths each year. Early diagnosis and monitoring the progression of the disease are crucial factors for the treatment and health indicators of a society. This study presents an innovative approach for the detection, segmentation, and classification of melanomas in dermoscopic images using advanced Computer Vision and Artificial Intelligence (AI) methods. Specifically, it applies Large Language Model (LLM) solutions for pre-diagnosis results through generative AI. This work explores combinations of methods for melanoma detection and segmentation based on the YOLO and SAM architectures, achieving 99% accuracy, surpassing various studies in the literature. The classification phase is based on a pipeline integrating feature extraction and selecting the best combination for melanoma region classification, achieving an accuracy of 86.0%, also outperforming different studies in the literature.

Palavras-chave: YOLO, Image segmentation, Accuracy, Generative AI, Malignant tumors, Large language models, Transfer learning, Pipelines, Feature extraction, Diseases