The Imaging and Machine Learning Research School (IMLRS) is hosted by the PhD School of Science, University of Copenhagen, and follows the regulations of the PhD school of Science. The IMLRS has an interdisciplinary focus, collaborating with most of the institutes at the faculty of Science, University of Copenhagen. There are five main research areas covered by the IMLRS, each using advanced computing capabilities to understand and solve complex problems.
Using medical imaging, new techniques for diagnosis, prognosis, monitoring and treatment are developed. Current research includes smokers lung characterization from CT scanning, atherosclerotic plague quantization in diabetes and aging from x-rays, osteoarthritis for knee MR scanning, and osteoporosis from lateral lumbar x-rays.
Machine learning is used to learn models and properties of natural phenomena from empirical data. As scientific data is becoming increasingly rich and plentiful, machine learning is becoming an important element in solving problems in a wealth of scientific disciplines. Interdisciplinary collaborations include physiotherapy, medicine, bioimaging, neuroscience, space science and oil companies.
High Performance Computing
Modern CPU architectures and parallel architectures provides new ways of harnessing computing powers, making way for emerging computing technologies such as multi core, GRID and GPU. The new computing power allows academic and industrial scientists to solve problems that previously have been too large or too time consuming to solve. This will be instrumental in finding cures for diseases, extracting information from satellites, exploring new planets and learning new details of nature and science. Applications include, bioinformatics, medical data analysis, molecular dynamics and protein ligand docking.
Computer vision embraces a variety of techniques for analysing still images and sequences. Among these are methods for extracting the meaningful structures, the information carried by a particular variation in the intensity or color pattern etc. The basic approach is to analyze the image data in a bottom-up fashion to free the encoded information. Computer vision and image analysis are closely related to research areas such as machine learning, signal processing and graphics.
Computer simulations are based on mathematical models of phenomena in areas such as physics, biology, economy etc. Simulations are used to predict real world behavior, allowing us to study and learn about the real world under idealized conditions which can not be achieved in a laboratory. Applications are wide, including nano-science, contact mechanics, robotics, bio-mechanics and computer animation.