Research Article Open Access

Error minimization in Brain tissue extraction for T1 weighted MR images

Abstract

In the current paper, a simple algorithm is proposed that employs an efficient and unique histogram analysis. A fundamental component of this analysis is an algorithm that partitions a histogram based on the position of the maximum deviation from a Gaussian fit. In the first step, the proposed partitioning algorithm is applied to delimit the highest peak in the gray-level histogram, aiming at identifying and discarding the background region. In the second step, Gaussian distribution is applied twice to the histogram of the head. The goal is to find two thresholds that encompass the gray-level range of the brain tissues and partition the histogram into two regions. A rough binary mask is then generated. Finally, a sequence of binary morphological operations is applied to the mask in order to completely isolate the brain. The significance of the work lies in the fact that it can successfully remove the unwanted regions found in Otsu’s method

G.L.N.Murthy, Dr.B.Anuradha, M.Siva Sankara Rao , K.L.Manassa

To read the full article Download Full Article