Two Lise 11 students recently placed first at the TÜBİTAK Science Competition in the Europe region, and in Turkey, with their paper “Early Diagnosis of Pneumonia with Deep Learning”, out of 17,000 projects presented. Prior to that, the students received the “Best Presenter Award” at the 2019 IEEE 4th International Conference on Computer and Communication Systems in Singapore in February 2019.
Using preprocessing techniques and neural networks on X-ray images, the students developed an algorithm for diagnosing pneumonia that has achieved 82.3% accuracy, the highest accuracy rate to date. They were also able to detect the region of pneumonia fluid from an x-ray image.
The students plan to work in cooperation with a research hospital to integrate their algorithm to the diagnosis process of actual patients.