WUT researchers create computer systems supporting medical diagnostics

“According to statistics, a radiologist spends nearly two hours analysing and describing the results of PET/MRI tests of a single patient,” says Tomasz Leś, DSc, from the Faculty of Electrical Engineering at WUT.

The systems on which WUT scientists are working are to improve the process. Based on the imaging data collections, which constitute real data (archived and anonymous), programmes for image processing devised by the Warsaw Medical University, Military Institute of Medicine, and Lower Silesian Oncology Centre are learning to detect lesions.

“The more images, including abnormal ones, are provided by the doctors, the better the tool we can devise to support their work,” says Dr Leś.

Thanks to artificial intelligence and analysis of numerous photos, the devices and their final application will facilitate automatic analysis of results, which will take a few minutes only. Additionally, these tools will allow us to analyse collectively the results obtained with different imaging methods and conduct multi-level diagnostics without focusing on the disease of one organ only.

International collaboration

“The project implementation involves the collaboration with the scientists from the universities in Barcelona and Italy, who deal with mathematical techniques of image processing and image quality enhancement, which is a great support for our research,” says Dr Leś. “Thanks to the work of foreign centres, we can achieve better results. The collaboration allows us to combine different methods and create larger systems, which later cooperate,” he adds.

-

Project “Development of machine learning methods to support medical diagnostics” is funded by the grant of the Research Centre POB Artificial Intelligence and Robotics within the Excellence Initiative – Research University programme implemented at the Warsaw University of Technology.

Research team:
Tomasz Leś, DSc; Żaneta Świderska-Chadaj, DSc; Krzysztof Siwek, PhD, DSc, Associate Professor; Professor Tomasz Markiewicz, PhD, DSc; Professor Stanisław Osowski, PhD, DSc; Estera Kot, MSc.