Abstract
In this paper, we introduce a novel automatic method for Corpus Callosum (CC) in midsagittal plane segmentation. The robust segmentation of CC in midsagittal plane is key role for quantitative study of structural features of CC associated with various neurological disorder such as epilepsy, autism, Alzheimer's disease, and so on. Our approach is based on Bayesian inference using sparse representation and multi-atlas voting which both methods are used in various medical imaging, and show outstanding performance. Prior information in the proposed Bayesian inference is obtained from probability map generated from multi-atlas voting. The probability map contains the information of shape and location of CC of target image. Likelihood in the proposed Bayesian inference is obtained from gamma distribution function, generated from reconstruction errors (or sparse representation error), which are calculated in sparse representation of target patch using foreground dictionary and background dictionary each. Unlike the usual sparse representation method, we added gradient magnitude and gradient direction information to the patches of dictionaries and target, which had better segmentation performance than when not added. We compared three main segmentation results as follow: (1) the joint label fusion (JLF) method which is state-of-art method in multi-atlas voting based segmentation for evaluation of our method; (2) prior information estimated from multi-atlas voting only; (3) likelihood estimated from comparison of the reconstruction errors from sparse representation error only; (4) the proposed Bayesian inference. The methods were evaluated using two data sets of T1-weighted images, which one data set consists of 100 normal young subjects and the other data set consist of 25 normal old subjects and 22 old subjects with heavy drinker. In both data sets, the proposed Bayesian inference method has significantly the best segmentation performance than using each method separately.
Highlights
The Corpus Callosum (CC), the largest white matter (WM) structure placed beneath the cortex of the human brain, connects homologous cortical areas of the two cerebral hemispheres (Caminiti et al, 2009)
The segmentation results of joint label fusion (JLF) was generated from optimal four primary parameters which are the radius of the image patch, the radius of the search neighborhood, the model parameter for transferring image similarity measures into atlas dependencies (β), and the weight of the conditioning identity matrix added to dependency matrix (α)
In OASIS data set, the results of mean Dice index are that JLF is 96.38 ± 1.78, Old likelihood estimated from sparse representation error (LESRE) is 92.96 ± 2.77 (%), LESRE is 94.09 ± 1.82 (%), prior information estimated from multi-atlas voting (PIEMV) is 95.44 ± 1.58 (%), and Bayesian inference based segmentation (BIbS) is 95.80 ± 1.52 (%)
Summary
The Corpus Callosum (CC), the largest white matter (WM) structure placed beneath the cortex of the human brain, connects homologous cortical areas of the two cerebral hemispheres (Caminiti et al, 2009). The CC defined in midsagittal plane (MSP) of the magnetic resonance imaging (MRI) has been generally used because it has relatively clear boundary line and mirrors overall properties of CC (Casanova et al, 2010; O’Dwyer et al, 2010; Joshi et al, 2013; Firat et al, 2014; Zhu et al, 2014; Balevich et al, 2015; Elahi et al, 2015; Wang et al, 2015). The manual segmentation of CC is a laborious and timeconsuming task The general shape of the CC is relatively regular, automated segmentation of CC has been still challenging because the detail shape is dynamic and the fornix, bundle of never fibers located around CC, has exceedingly similar intensity as CC
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
