In St. Petersburg, artificial intelligence is taught to identify ulcers and cancer from a pic

Experts from Peter the Great St. Petersburg University are teaching artificial intelligence to detect stomach ulcers, cancer and gynecological diseases from a photo. 

In early 2021, it is planned to conduct the first clinical trials of the technology in the laboratory of the Almazov National Medical Center. Image via: pixabay.com During the research, doctors inject plastic tubes of endoscopes into the human body in order to film the processes taking place in the digestive system. To analyze the images, artificial intelligence (AI) was used, the database of which was filled with information about subjects with healthy and diseased gastrointestinal tract. Further, the computer complex will compare the images from the database with the photos obtained during the examination of the patient. 

If there are deviations, doctors will receive appropriate signals and preliminary diagnoses. It is assumed that AI will timely notice gynecological diseases, stomach ulcers and malignant tumors. The operation of the computer complex is divided into three stages: detection, segmentation and classification. First, the AI ​​detects the lesion of the mucous membranes, then the image is divided into parts with an indication of the painful lesions, after which the system determines whether the deviation matches the photo of known problems. 

Scientists have two important tasks: to obtain high-quality images and to make the system accurately interpret the data based on the analysis of visual signs. Currently, machine analysis is doing a good job with the responsibilities assigned to it.