APPLICATION OF ARTIFICIAL INTELLIGENCE IN IMAGE DIAGNOSIS

• DOI: 10.22533/at.ed.3173212313061

• Palavras-chave: Artificial intelligence; Imaging diagnosis; Algorithms; Deep learning.

• Keywords: Artificial intelligence; Imaging diagnosis; Algorithms; Deep learning.

• Abstract:

  Artificial intelligence (AI) is human-like intelligence triggered by software. It is a set of complex mathematical models based on the structure and functioning of biological neurons. The present work aims to present, based on the scientific literature, the fundamentals as well as the applications of AI in diagnostic imaging. Among the different AI tools with potential to aid diagnostic imaging, computer-aided diagnosis (CAD) and the convolutional neural network (CNN) stand out. CAD is a computational system that uses results from automated quantitative analysis of radiographic images recorded in a database. The use of CAD aims to verify the radiologist's interpretation and improve the accuracy of the imaging diagnosis, using the computer response as a reference. As contributions of this tool, one can cite the aid to image processing through a computational system containing a database with patterns considered normal and abnormal. CNNs, on the other hand, are capable of identifying molecules with potential in the treatment of cancer and interpreting computed tomography images using a worldwide database of images associated with typical diagnostic terms. The use of AI as a diagnostic aid aims to propose a system that can be implemented in traditional imaging exams. Although still incipient, the expectation is that we will experience advances in the area of diagnostic imaging as new AI algorithms are developed for this purpose.