tissues in the brain separated by an interface. Extraction of this boundary is very important for quantitative analysis and monitoring of atrophy. A perfect capture for this purpose is a missing vital item for which research continues. In order to get close to such precision, two novel systems are presented for segmentation of cortical GM and WM in brain MRIs. The system imitates human perceptual sensitiveness to contrast, which is the basic principle of edge detection algorithms, and completes its shortcoming feature of segmentation. As well as being completely automatic, the system is also unsupervised. The correct GM-WM boundary rate gets close to 77{\%} and the segmentation accuracy is over 95{\%}, which are promising results. In comparison tests with SPM, the proposed technique showed 6{\%} higher accuracy with both noisy and normal data and better recovery of small cavities in sulci, being confirmed by experts' drawings.