A project dedicated to training artificial intelligence to analyze tomographic images has been launched at Novosibirsk State University (NSU). The initiative aims to support doctors in diagnosing oncological diseases and developing new drugs.
The work is being carried out by the Laboratory of Innovative and Nuclear Medicine of the Faculty of Physics at NSU. It is implementing AI technologies for analyzing data obtained using computed tomography. In addition, methods of neutron capture therapy are being developed and tests of new drugs for cancer treatment are being conducted.
Physicists at NSU aim to train AI to perform routine and time-consuming tasks, such as interpreting tomographic images. For this purpose, a tomographic atlas of normal and pathological conditions of animals is being created, which will be used to train neural networks. According to Vladimir Kanygin, head of the laboratory, AI will be trained in three areas: medicine, veterinary medicine, and pharmacology.
Ulyana Krechetova, a member of the research group, noted that transferring routine tasks to AI will increase overall efficiency. Describing images on CT scans requires a lot of time and high concentration, and can also vary depending on the specialist. Using AI will help reduce the influence of the human factor, such as fatigue and distraction, and in some cases will provide more accurate results.
Krechetova and her graduate students have already achieved initial success in image segmentation, highlighting organ contours. The immediate plans include developing models and programs for classifying tumors and helping doctors in identifying pathologies.
Read more on the topic:
Russia wants to train AI to create cancer vaccines