Um sistema evolutivo para auxiliar o diagnóstico da cifose e lordose

Abstract

This work proposes the implementation of a variation of Holland's genetic algorithm to monitor the evolution of a patient's posture through the analysis of exams performed by the patient with the aid of a mobile device. The purpose of this system is to help the patient to identify the presence of changes in the spine called kyphosis and lordosis, at which time the patient will need to seek medical help. Therefore, the system must receive information about the limits of the spine angles considered normal, acceptable and critical for kyphosis and lordosis. Then, the system will monthly assist in making the diagnosis, generated from a photo of the patient's profile taken by a mobile device, from which a set of information is extracted, which is recorded in a database, to be analyzed by a genetic algorithm. The genetic algorithm generates several evolution possibilities for the patient's spine from the information recorded in the database, if this information leads the genetic algorithm to find a situation considered critical for the patient in the coming months, the system will display a message on the screen cell phone informing the patient of the need to see a doctor to have their spine evaluated. To validate the implementation of this system, simulations were carried out with images generated with photoshop. The results obtained with these simulations demonstrate the system's ability to predict in advance changes in the patient's posture that can lead to kyphosis or lordosis.