Localization of Foci Pathological Radiopharmaceutical Accumulation Based on Artificial Intelligence Methods

This paper deals with the technology of localization of pathological foci of radiopharmaceutical accumulation based on deep learning methods. It can be used in departments of radionuclide diagnostics (nuclear medicine) as a second opinion in the description of radionuclide studies by a radiologist, as well as a way of early diagnosis of tumors and metastatic lesions, assessment of the effectiveness of treatment, including differential diagnosis of the disease. The paper presents the results of computational experiment conducted with the trained model on test data. As a result, the model achieved an accuracy of 89% on the test sample.