One of the most terrible threats of coronavirus infection is the difficulty of timely diagnosis. Due to the possibility of a symptomless course of the disease, it is possible to detect the disease in time only with the help of computer tomography. However, the doctor may not notice tissue damage in the early stages.
That is why the creation of accurate detection models can help automate the examination and detect potential threats that a live specialist may not notice. In this dataset, we selected various projections of lung images and enlisted the advice of doctors who helped to assess the accuracy of the markup.