Inactive

The research activities of the Quantitative Medical Imaging Section, NIBIB, focus on developing quantitative Magnetic Resonance Imaging (MRI) techniques with the goal of obtaining clinically informati... The research activities of the Quantitative Medical Imaging Section, NIBIB, focus on developing quantitative Magnetic Resonance Imaging (MRI) techniques with the goal of obtaining clinically informative biomarkers that are validated in animal models and human subject studies. One of the most promising approaches that our Section has pioneered is the use of diffusion MRI to achieve these goals. Diffusion MRI data is very large in size and requires extensive processing computations to be quantitatively viable. QMI processes and analyzes large quantities of medical imaging data, mostly MRI of human and animal brains using different imaging modalities. These data need to be preprocessed to correct for imaging artifacts and distortions and post-processed to account for variability among subjects in order to achieve robust and reliable conclusions. The processing and analysis software tools that are being developed by the section as a part of the project, which are also publicly available, can utilize the entire processing power provided by CUDA-enabled video cards. Therefore, these powerful video cards will be used for data processing & analysis, software development. Additionally, the Section has been developing machine-learning based tools for both MRI data processing and analysis. The requested DGX A100 system is specially optimized for such tasks.