TY - GEN
T1 - Novel stochastic framework for accurate segmentation of prostate in dynamic contrast enhanced MRI
AU - Firjany, Ahmad
AU - Elnakib, Ahmed
AU - El-Baz, Ayman
AU - Gimel'farb, Georgy
AU - El-Ghar, Mohamed Abo
AU - Elmagharby, Adel
PY - 2010
Y1 - 2010
N2 - Prostate segmentation is an essential step in developing any non-invasive Computer-Assisted Diagnostic (CAD) system for the early diagnosis of prostate cancer using Dynamic Contrast Enhancement Magnetic Resonance Images (DCE-MRI). In this paper we propose a novel approach for segmenting the prostate region from DCE-MRI based on using a graph cut framework to optimize a new energy function consists of three descriptors: (i) 1st -order visual appearance descriptors of the DCE-MRI; (ii) a spatially invariant 2nd -order homogeneity descriptor, and (iii) a prostate shape descriptor. The shape prior is learned from a subset of co-aligned training images. The visual appearances are described with marginal gray level distributions obtained by separating their mixture over the image. The spatial interactions between the prostate pixels are modeled by a 2nd -order translation and rotation invariant Markov-Gibbs random field of object / background labels with analytically estimated potentials. Experiments with prostate DCE-MR images confirm robustness and accuracy of the proposed approach.
AB - Prostate segmentation is an essential step in developing any non-invasive Computer-Assisted Diagnostic (CAD) system for the early diagnosis of prostate cancer using Dynamic Contrast Enhancement Magnetic Resonance Images (DCE-MRI). In this paper we propose a novel approach for segmenting the prostate region from DCE-MRI based on using a graph cut framework to optimize a new energy function consists of three descriptors: (i) 1st -order visual appearance descriptors of the DCE-MRI; (ii) a spatially invariant 2nd -order homogeneity descriptor, and (iii) a prostate shape descriptor. The shape prior is learned from a subset of co-aligned training images. The visual appearances are described with marginal gray level distributions obtained by separating their mixture over the image. The spatial interactions between the prostate pixels are modeled by a 2nd -order translation and rotation invariant Markov-Gibbs random field of object / background labels with analytically estimated potentials. Experiments with prostate DCE-MR images confirm robustness and accuracy of the proposed approach.
UR - http://www.scopus.com/inward/record.url?scp=78149387975&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78149387975&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-15989-3_14
DO - 10.1007/978-3-642-15989-3_14
M3 - Conference contribution
AN - SCOPUS:78149387975
SN - 3642159885
SN - 9783642159886
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 121
EP - 130
BT - Prostate Cancer Imaging
T2 - International Workshop on Prostate Cancer Imaging: Computer-Aided Diagnosis, Prognosis, and Intervention Held in Conjunction with MICCAI 2010
Y2 - 24 September 2010 through 24 September 2010
ER -