Title :
Segmentation of MRI brain with a bootstrapped version of the HMRF-EM algorithm
Author :
Mabrouk, Sabra ; Mhiri, Slim ; Ghorbel, Faouzi
Author_Institution :
CRISTAL Lab., Nat. Sch. of Comput. Sci., La Manouba, Tunisia
Abstract :
We are interested in this work to optimize the algorithmic complexity of Markovian segmentation of brain tissues in MRI by the Bootstrap sampling. The introduction of this resampling allows to create the independence conditions which gives a better convergence of the mixture identification algorithm. A comparative study is made with the non-bootstrapped version in the mean of the misclassification rate.
Keywords :
biological tissues; biomedical MRI; expectation-maximisation algorithm; image classification; image segmentation; medical image processing; optimisation; HMRF-EM algorithm; MRI brain segmentation; Markovian segmentation; algorithmic complexity; bootstrapped version; brain tissues; convergence; expectation-maximization algorithm; misclassification rate; mixture identification algorithm; optimization; Hidden Markov models; Image segmentation; Magnetic resonance imaging; Noise; Parameter estimation; Radio frequency; Robustness;
Conference_Titel :
Electrotechnical Conference (MELECON), 2012 16th IEEE Mediterranean
Conference_Location :
Yasmine Hammamet
Print_ISBN :
978-1-4673-0782-6
DOI :
10.1109/MELCON.2012.6196476
