UM ESTUDO SOBRE A UTILIZAÇÃO DE INTELIGÊNCIA ARTIFICIAL NA DETECÇÃO DE CÂNCER DE PELE

Abstract

Artificial intelligence (AI) emerges as a technology with capacity to revolutionize several fields, including diagnostic medicine. Through a bibliographic review, this work explored the definition and applications of AI, highlighting its use in detecting skin cancer, using databases such as CAPES and PubMed to select books, articles and theses relevant to the topic. Traditional diagnostic methods were analyzed in conjunction with advanced architectures, such as convolutional neural networks (CNNs) and Generative Adversarial Networks (GANs), which expand diagnostic possibilities. Moleanalyzer Pro analysis demonstrated that CNNs provide comparable performance to dermatologists, although professionals still outperform algorithms when they have access to detailed clinical information. Despite their high performance, CNNs face limitations, especially in complex cases such as melanomas in specific regions. It is concluded that the impact of AI on the agility of diagnosis is significant, as the automation of image analysis can speed up the process and enable earlier diagnoses, especially in contexts with a shortage of professionals. However, challenges such as lack of data and generalization of diagnoses still need to be addressed. Keeping the investment in AI in diagnostic medicine is crucial, especially considering the high incidence of skin cancer in Brazil, since its early detection is vital for effective treatment.