Speeding up HMRF_EM algorithms for fast unsupervised image segmentation by Bootstrap resampling: Application to the brain tissue segmentation

M'Hiri, Slim; Cammoun, Leila; Ghorbel, Faouzi

doi:10.1016/j.sigpro.2007.04.010

Cited by 17 publications

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A bootstrap sample is obtained by independent drawings with replacement from the empirical distribution (given by data histogram H ). The question of finding an optimal number M = card {A} of representative data has been addressed in [1] (see also [21]): if G denotes the number of bins in H then M is the highest value such that T (M) < with…”

Section: Subsampling-based Parameter Estimationmentioning

confidence: 99%

Exact Fast Computation of Optimal Filter in Gaussian Switching Linear Systems

Derrode

Pieczynski

2013

IEEE Signal Process. Lett.

View full text Add to dashboard Cite

This paper presents a contextual algorithm for the approximation of Baum's forward and backward probabilities, which are extensively used in the framework of Hidden Markov chain models for parameter estimation. The method differs from the original algorithm by taking into account only a neighborhood of limited length and not all the data in the chain for computations. It then becomes possible to propose a bootstrap subsampling strategy for the computation of forward and backward probabilities, which greatly reduces computation time and memory saving required for EM-based parameter estimation. Comparative experiments regarding the neighborhood size and the bootstrap sample size are conducted by mean of unsupervised classification error rates. Practical interest of such an algorithm is then illustrated through the segmentation of large-size images; classification results confirm the validity and the accuracy of the proposed algorithm while greatly reducing computation and memory requirements.

show abstract

Section: Subsampling-based Parameter Estimationmentioning

confidence: 99%

Exact Fast Computation of Optimal Filter in Gaussian Switching Linear Systems

Derrode

Pieczynski

2013

IEEE Signal Process. Lett.

View full text Add to dashboard Cite

show abstract

“…This work [23] deals with global statistical unsupervised segmentation algorithms. In the context of MRI, an accurate and robust segmentation can be achieved by combining both the Hidden Markov Random Field (HMRF) model and the Expectation-Maximization (EM) algorithm.…”

Section: Review Of Previous Work On Brain Mri Segmentationmentioning

confidence: 99%

Morphological Segmentation Based Fuzzy Features for Retrieval of Brain MRI

Ingole¹,

Kulat²

2011

IETE J Res

View full text Add to dashboard Cite

Retrieval of domain-specific images is an important research area. Morphological skull stripping and segmentation is proposed for highlighting and extraction of the region-based features of human brain T 2 -weighted magnetic resonance (MR) images. Further, fuzzy representation of these features and its use in retrieval of brain MR imaging (MRI) is demonstrated in this paper. Unified Fuzzy Matching technique is used to assess the image-to-image similarity of MRI. Segmentation results show a marked improvement in the quality of segmentation as compared to fuzzy c-means clustering method and it is confirmed with manual segmentation. Further, the use of features obtained with this technique in retrieval has also been found to give better retrieval results in terms of precision, average rank and mean standard deviation. Such brain MRI retrieval system is useful tool to assist doctors in diagnosis by quickly searching similar cases in this domain.

show abstract

“…• Estimation of the distribution. A Bootstrap resampling can be added to the HMRF-EM framework in order to speed up the classification process [8].…”

Section: B Hmrf-em With Bootstrap Resamplingmentioning

confidence: 99%

“…This statistical classification method is based on Bayesian analysis and performed with EM algorithm and HMRF modeling used to encode spatial information through the mutual influences of neighboring voxels for class assignments [15]. A Bootstrap resampling [1], [5] is used to accelerate classification process [8].…”

Section: Introductionmentioning

confidence: 99%

Segmentation of liver tumor using HMRF-EM algorithm with Bootstrap resampling

Said¹,

Azaiz²

2010

2010 5th International Symposium on I/v Communications and Mobile Network

View full text Add to dashboard Cite

Volume measurement of liver tumor is an important task for surgical planning and cancer following-up. The computation of this volume requires an efficient liver tumor segmentation method. This work deals with liver tumor segmentation from computed tomography (CT) images. We are interested by HMRF-EM classification method. This method considers the spatial information given by voxel neighbors. A Bootstrap resampling, based on selecting randomly an optimal representative set of voxels, is proposed to accelerate the classification process. In order to extract correctly the tumor region, a post-treatment based on morphological operators and active contours method is needed. The entire approach was evaluated on two clinical data sets with manually generated ground truth segmentation by radiologists and has presented promising results.

show abstract

Speeding up HMRF_EM algorithms for fast unsupervised image segmentation by Bootstrap resampling: Application to the brain tissue segmentation

Cited by 17 publications

References 17 publications

Exact Fast Computation of Optimal Filter in Gaussian Switching Linear Systems

Exact Fast Computation of Optimal Filter in Gaussian Switching Linear Systems

Morphological Segmentation Based Fuzzy Features for Retrieval of Brain MRI

Segmentation of liver tumor using HMRF-EM algorithm with Bootstrap resampling

Contact Info

Product

Resources

About