Strabismus is an ophthalmological disease characterized by an imbalance in the ocular muscles, so that the two eyes do not fix the same point or object at the same time. Its incidence can occur from the first months of life to adulthood. Late detection may cause double vision or loss of vision, hence the importance of early detection of strabismus for its prognosis and effective treatment. That said, in this work is proposed a new method that uses neural networks to automatically detect strabismus by evaluating the positioning of the limbus center in relation to the center of the corners of the eyes. To do so, we use techniques of image processing and machine learning, obtaining sensitivity of 66,7%, specificity of 60% and accuracy of 65,6%. Resumo: O estrabismoé uma doença oftalmológica caracterizada por um desequilíbrio nos músculos oculares, fazendo com que os dois olhos não fixem o mesmo ponto ou objeto ao mesmo tempo. Sua incidência pode ocorrer desde os primeiros meses de vida até a idade adulta. A detecção tardia pode causar visão dupla ou perda da visão, por isso a importância da detecção precoce do estrabismo para seu prognóstico e tratamento eficaz. Dito isso, neste trabalhoé proposto um novo método que utiliza redes neurais para detectar automaticamente o estrabismo avaliando o posicionamento do centro do limbo em relação ao centro dos cantos dos olhos. Para tanto, utiliza-se técnicas de processamento de imagens e aprendizado de máquina, obtendo sensibilidade de 66,7%, especificidade de 60% e acurácia de 65,6%.
Strabismus is a pathology that affects the parallelism between two eyes. Estimates show that 1.3% to 3.5% of children aged six months to six years have strabismus around the world. Strabismus can lead to irreversible loss of vision, so early detection and appropriate treatment increase the likelihood of alignment being restored to normal. That said, in this study we proposed an automatic evaluation system to detect strabismus in face images, analyzing the horizontal positioning of the limb center in relation to the center of the corners of the eyes. The proposed method consists of five steps: (1) acquisition, (2) detection of the eye region, (3) segmentation and reconstruction of the sclera, (4) detection of the limb region and corners of the eyes and (5) identification of the presence of strabismus, with this approach we obtain 90.1% sensitivity, 100% specificity and 91.1% accuracy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.