Automatic Segmentation of Infant Brain MRI using Soft Computing Techniques

This article is concerned with exploration and diagnostic implementation of an effective neo-anatomical brain MRI classification method to classify primal cognitive development and investigate neuro-anatomical intellectual disability correlations. A crucial stage in the research as well as appraisal of the newborn brain growth is neonatal brain tissue classification. Owing to the major variations in anatomy and tissues among neonate and mature brains, the largest proportion of developing technology for the classification and segmentation of the adult brain really aren't sufficient for newborns brain. The existing brain tissue classification strategies for MRIs rely either on manual interactions or involve the use of atlases or models, which ultimately skew the findings from the population used to extract atlas. This article, focuses on atlas free soft computing approach to classify the neonatal brain tissue. Classification of brain tissue is the main process in which regional brain tissue examination is conducted. This helps the regional brain development to be characterized and the correspondence with therapeutic conditions to be studied. The modified BM3D approach is utilized for image enhancement along with 32 Gabor filter bank-based feature extraction. The innovative aspect of this research is the multistage classification methodology, which produces higher dice coefficients and lower MHD values when compared to existing approaches