About

    This is a test project for the recognition of X-ray images of patients with COVID-19 pneumonia. The use of a neural network here provides the correct result in 92-96% in the test data set.

    The neural model is now set up for processing ONLY CHEST X-ray images!

    Uploading OTHER images will give INCORRECT answers.

    In total, for training and testing the network, it was used about 5,000 chest X-rays of patients diagnosed with Covid-19 and patients without this diagnosis.

The project structure included the following elements:

    1) Formation, markup and normalization of training, test and verification data sets.

    2) Creating a neural network model.

    3) Training and retraining of the neural network model to ensure an acceptable level of accuracy.

    4) Checking and calibrating the model based on test and verification data.

    5) Application of genetic algorithms to improve the structure of the neural network model.

    6) Writing neural model logic by python.

    7) deploying the application (Django framework) with a working neural model (API for processing incoming data and return of processing results) with initial settings on the server.