Automated skull stripping in brain MR images

Skull stripping is a significant as well as a preliminary step in diagnosing brain disorders. It removes extra-meningeal tissues from Magnetic Resonance Images of the brain. Magnetic Resonance Imaging (MRI) is a widely used technique for analysis of brain images. An efficient hardware-based algorithm for Skull segmentation would help in developing an automated brain image analysis system for real time applications in biomedical sciences. In this work, a Raspberry Pi single board computer based image analysis algorithm for an automatic skull stripping is reported. The experiment has been carried out with T1 weighted axis images. In order to reduce the noise and enhance the quality, initially the images were pre-processed. Further, edge detection and morphological operations were performed to extract the skull from the brain images. The proposed method has been validated by evaluating quantitative performance metrics like Jaccard similarity index and the Dice coefficient. This technique will serve as the major step for technological outbreaks for developing systems for automated skull stripping the images of the brain in the future.