Presenting our winning Machine Learning solution at the renowned SPIE Medical Imaging conference

17. februar 2022 - Avtorja Jasna Simončič, Lucija Korbar

17. februar 2022
Avtorja Jasna Simončič, Lucija Korbar

By being a part of numerous research projects, we constantly deepen our knowledge, enrich our products and solutions with gained expertise, and, more importantly, contribute to technological advancement in a meaningful way. Our experts continuously prove they’re the best in their dedicated fields, which was again the case a few months ago, when our AI team has won the SegPC-2021 Plasma Cell Segmentation competition with their machine learning expertise, which earned them a spot among world’s leading researchers, contributing to the advancement of medical imaging technologies. Our AI experts will present their winning solution at the renowned SPIE Medical Imaging conference, happening 20 - 24 February 2022 in San Diego!

 

Best machine learning algorithm for cell segmentation

Our experts are distinguished by their ability to skillfully apply expertise to many different fields. Last year, our AI team decided to put their data science skills to good use and be part of the development of a breakthrough solution, that helps diagnose different types of cancer using computer-assisted tools. Cancer is usually diagnosed by performing a tissue biopsy, but the use of computer-assisted tools could be a real game-changer in terms of accuracy and speed. As part of the iPC project, the team entered the SegPC-2021 challenge for plasma cell segmentation, organized as part of the IEEE International Symposium on Biomedical Imaging 2021. The goal was to develop a method for instance segmentation of Multiple Myeloma plasma cells.

The competition was fierce, but our team did an amazing job. Among 41 competing teams from around the world, our machine learning algorithm stood out the most. Read more about the challenge and the wining solution here. And now they will present the solution at the SPIE Medical Imaging conference, the event where the science of medical imaging is explored.

 

More achievements of our AI team

After this amazing achievement, the team didn’t just sit back and rest on its laurels for the rest of the year. Quite the opposite, actually! In June, at the CVMI 2021 workshop that was organized as part of the well-established IEEE International Conference on Computer Vision and Pattern Recognition, our colleague Dejan presented his unsupervised method for distinguishing cancerous from healthy tissue, that almost matches the performance of supervised learning.

A panel of internationally respected experts recognized the work on Constellation-Based Deep Ear Recognition by Dejan and his co-authors as a significant contribution to biometric research and innovation in Europe and the work received the prestigious European Biometrics Max Snijdr Award.

And last but not least, our colleague Tomaž and his team participated in The Data Science Project Competition, organized annually by the Faculty of Computer and Information Science, University of Ljubljana, where they leveraged deep learning and graph neural network approaches to develop an effective recommendation system suggesting relevant legal documents to a user from tens of thousands of documents, based on various factors. Their solution was awarded 2nd place out of 11 teams and 5 finalists.

Undoubtedly, 2021 was a very successful year, thanks to our colleagues who eagerly took advantage of these opportunities, shared their knowledge in various fields and left a significant mark! Congrats to the entire team and good luck with the new challenges!

  This work is partially funded by European Commission project iPC (grant agreement number 826121).