Swift Cerebral Modelling with DenseUNet

Quantitative and volumetric investigations utilizing magnetic resonance imaging are an essential technique used by researchers to examine the brain throughout normal maturation and different neurological illnesses, including aging. Despite the processing time and need for human adjustment, volumetric analysis is still commonly employed in clinical practice despite the unavailability of open-source brain segmentation tools. Here, we provide a comprehensive segmentation of deep brain and cortical gray matter structures that match typical aseg+aparc areas in the popular open-source software Freesurfer. We are extending the usage of machine learning models beyond proof-of-con, as was previously announced. This study aids clinical translation and quantitative brain segmentation research by offering a quick and easy-to-use deep learning-based brain segmentation tool. Improved segmentation quality and a shorter processing time are two benefits of the offered technologies. This is the first research to apply a surface-based approach, which is widely used by specialists in the area, to swiftly and precisely segment 102 brain regions. Additionally, this is the first time that knowledgeable readers have tried to evaluate the level of segmentation quality attained using deep learning-based models. We show that our deep learning based model outperforms Freesurfer, a conventional segmentation tool. DARTS is the name of the suggested deep learning-based technology.