Exploring the utility of new Artificial Intelligence and Deep Learning methods for radiological image analysis and lung disease diagnosis

Current state of collaboration.

The MedGIFT research group of the University of Applied Sciences Western Switzerland (HES–SO), Sierre, Switzerland (Swiss group) and the Biomedical Image Analysis group from United Institute of Informatics Problems, National Academy of Sciences of Belarus (Belarus group) have had a remote collaboration based on emails, telephone conferences and common conference organizations for over three years. The major points of mutual interest include development of new methods and software for biomedical image analyses and computer assisted diagnosis.

The project of this current application aims at performing the following collaborative activities:

1. Exchange of open biomedical image data (at least 1000 Computed Tomography chest scans and Chest X-rays).

2. Carrying out a workshop in Switzerland. Giving presentations on novel Artificial Intelligence and Deep Learning methods for radiological image analysis and lung disease diagnosis, focusing on tuberculosis and lung cancer.

3. Development and approval of study protocols for benchmarking and comparison of the effectiveness of methods and software developed by both parties.

4. Organizing and conducting the ImageCLEF 2019 international benchmarking campaign on multimodal data analysis methods, mainly for radiological images and lung diseases.Carrying out a workshop in Belarus.

5. Performing analytical comparison of the results on image analysis and lung disease diagnosis achieved by Swiss and Belarusian teams. Scoring and publishing of the ImageCLEF 2019 results.

The above activities stick to the Open Science philosophy in the biomedical domain. It is planned to create a solid ground for future European H2020 project submissions and possible other joint project applications and collaborations.

Participants: