Development of a Machine Learning based model for early screening for oral cancer

• Autor(a) principal: Pereira da Silva, Ivisson Alexandre
• Data de Publicação: 2023
• Outros Autores: Rodrigues Rosa de Oliveira, Catarina, dos Santos Oliveira, José Marcos, Correia Lessa Filho, Carlos Alberto, Soares Ferreira, Sonia Maria
• Tipo de documento: Artigo
• Idioma: por; eng
• Título da fonte: Diversitas Journal
• Texto Completo: https://diversitasjournal.com.br/diversitas_journal/article/view/2532
• topic: Câncer de boca; Inteligência artificial; Machine learning; Oral cancer; Artificial intelligence; Machine learning

Oral Squamous Cell Carcinoma (OSCC) is the most frequent type of oral cancer, accounting for about 40% of malignant head and neck lesions. It ́s known that the favorable prognosis is associated with early diagnosis, since the survival rate increases as a function of the diagnosis in the early stages of the disease. Thus, the objective of this work was to implement and train a Machine Learning model that can help in the diagnosis of oral cancer. Through technologies such as artificial intelligence (AI) that can use images in their analyses, it ́s sought to improve the prognosis of oral cancer through its early detection. Using the branch of AI, Machine Learning and its subgroup Deep Learning, it becomes possible through Convolutional Neural Network (CNN) to perform an image screening of malignant and premalignant lesions, in order to identify the presence or not of oral cancer. The RNC structure is based on the MobileNet structure, which separates the images into fragments and after training, showed the identification of cancer in 91% of the images examined and of Leukoplakia in 84% of the analyzed images.