The Knowledge Engineering Institute (IIC) has carried out research that analyzes the applicability of Deep Learning to distinguish between x-rays of healthy patients, patients with pneumonia and those who present pneumonia as a result of COVID-19.
The main objective of the research is to train a convolutional network with X-ray images of three types of patients: without any lung disease (Normal), with symptoms of COVID-19 and people with symptoms of pneumonia not related to COVID-19.
And it is that Machine Learning and Deep Learning systems need to learn from data, in this case images, that reflect what they will later have to detect automatically. For the development of this study, 3 data sources were used. After selecting the data, the Deep Learning model is developed and trained with the examples, validating its progress with new data. Its effectiveness is then measured to classify X-rays with a set of images that have not been used in training the network.
When testing the new model with the test data, an accuracy of more than 94% was obtained in the detection of COVID-19 in radiographs.
This research is another example of the importance of Artificial Intelligence to detect this disease early and effectively and apply the relevant solutions.
https://www.iic.uam.es/lasalud/deteccion-covid19-en-radiografias-con-deep-learning/
