Can Algorithms Used in Automatic Preprocessing Affect the Final Value of Brain Volume? - Preliminary Study
摘要
This paper focuses on the role of MRI research and postprocessing of images. The main aim of the study is to present comparative analyzes of the brain using magnetic resonance imaging, taking into account volumetric calculations and tools for automatic detection and segmentation of demyelinating changes, especially in terms of the diagnosis of multiple sclerosis or Alzheimer's disease. Additional parameters were calculated, such as the parenchymal fraction factor (BPF), which allows for indepth diagnostics and precise monitoring of drug therapies. BPF depends on many factors, however, the work focuses on the calculation method. The subject of pre and postprocessing is one of the most important elements of modern medicine, because diagnostics and diagnosis are based on analyzes. Attention should be paid to the pre- and postprocessing as the volume and BPF values for 9 different analyzes are different. Normalization seems to help solve some of the problems and DPARSF or CAT poses a suitable alternative with white and gray matter segmentations.