MDLMA (“Multi-task Deep Learning for Large Multimodal Biomedical Image Analysis”) is a joint project of the Helmholtz Center Hereon, the German Electron Synchrotron DESY, the University of Lübeck (UzL) and the company Syntellix. It is funded by the Federal Ministry of Education and Research (BMBF), funding number 031L0202A. The project is led, managed and coordinated by Hereon.
Note: This site focusses on DESYs contributions to the MDLMA project
In order to optimize Mg-based biodegradable implants in terms of their mechanical, biological and degradative properties, to understand the mechanisms of interactions between microstructure, mechanical properties, biology and degradation, and to tailor implants for specific applications, large amounts of multimodal biomedical image data have to be analyzed become. The modalities include laboratory-based X-ray computed tomography (CT), synchrotron radiation-based microcomputed tomography (SRμCT), magnetic resonance imaging (MRI), small-angle X-ray scattering (SAXS), and histology. The common image analysis tasks considered here include registration, segmentation, or segmentation. Classification and image enhancement (e.g. artifact or noise reduction) using deep learning (DL) approaches. In addition, new multi-task DL methods are being developed so that various complementary tasks can be holistically combined and information about the individual analysis tasks can be transferred. In order to enable the application of these multi-task solutions for other areas, modalities and tasks as well, a unified software platform is being developed that enables fast and efficient implementation and application of data analysis tasks.
The pages are static autogenerated by CI/CD pipelines at gitlab.desy.de.
The Federal Ministry of Education and Research (BMBF) is funding the MDLMA project (031L0202C) within the Computational Life Sciences funding measure as part of the federal research program on Digitalisation and Artificial Intelligence.