Machine learning for automated analysis of radiology examinations
Machine learning and computer vision is gaining traction in clinical radiology. We are looking for students to participate in model development and building of different automated pipelines for clinical radiology tasks.
The aim is to develop machine learning models and pipelines for automated clinical radiology tasks.
Machine learning and computer vision is being used more and more extensively for solving various radiology tasks. Nevertheless, clinical radiology is still heavily dependent on manual inspection of images. Automation of steps in the image review process can contribute to e.g. more efficient prioritization of cases for the radiologist’s review, and shorter time to diagnosis for the patient. There are however great demands on such tools to be able to use them in the clinical setting.
In this project, we are developing machine learning models and pipelines for automatic detection or grading of common conditions, such as knee osteoarthritis and pulmonary embolism. For the work, several large data sets of X-ray and CT images from e.g. lungs and knees have been annotated by our collaborating radiologists.
We are looking for students to participate in model development and building of different automated pipelines for clinical radiology tasks. Knowledge of the basics of machine learning and some basic programming skills are required. Experience of using Python for data science and/or machine learning is an advantage.
Professor Tobias Sjöblom
Read more about our research here.