Authors: Shoaib Amin Banday; Ajaz Hussain Mir
Addresses: Department of Electronics and Communication Engineering, National Institute of Technology, Hazratbal, Srinagar J&K, India ' Department of Electronics and Communication Engineering, National Institute of Technology, Hazratbal, Srinagar J&K, India
Abstract: The work herein proposes a framework for semi-automatic segmentation of pituitary gland from MRI brain images. The proposed framework initially uses a fused stationary wavelet transform (SWT) and discrete wavelet transform (DWT) to obtain a high resolution image of the input MRI brain image. After the input MRI brain image enhancement, the method applies thresholding and mathematical morphology to segment the pituitary gland from the input MRI brain image. The proposed algorithm for the same is coded in MATLAB 7.9 on MRI brain images. The segmented pituitary gland obtained using the proposed method is compared with the manually segmented pituitary gland (by an expert), region growing-based brain segmentation and watershed brain segmentation using Jackard's similarity coefficient (JSI) and overlap index (OI). The visual evaluation by a team of radiologists has demonstrated the efficacy of the proposed framework of pituitary gland extraction.
Keywords: pituitary gland; brain segmentation; magnetic resonance imaging; MRI; stationary wavelet transform; SWT; discrete wavelet transform; DWT.
International Journal of Medical Engineering and Informatics, 2017 Vol.9 No.3, pp.201 - 219
Available online: 10 Apr 2017 *Full-text access for editors Access for subscribers Purchase this article Comment on this article