TY - GEN
T1 - A novel 3D joint MGRF framework for precise lung segmentation
AU - Abdollahi, Behnoush
AU - Soliman, Ahmed
AU - Civelek, A. C.
AU - Li, X. F.
AU - Gimel'farb, G.
AU - El-Baz, Ayman
PY - 2012
Y1 - 2012
N2 - A new framework implemented on NVIDIA Graphics Processing Units (GPU) using CUDA for the precise segmentation of lung tissues from Computed Tomography (CT) is proposed. The CT images, Gaussian Scale Space (GSS) data generation using Gaussian Kernels (GKs), and desired maps of regions (lung and the other chest tissues) are described by a joint Markov-Gibbs Random Field (MGRF) model of independent image signals and interdependent region labels implemented on GPU. The initial segmentation from the original and the generated GSS CT images is based on the Linear Combination of Discrete Gaussian (LCDG) models; The initial segmentation is obtained from the original and the generated GSS CT images; then they are iteratively refined using a parallel MGRF model implemented on GPU with analytically estimated potentials. Finally, these initial segmentations are fused together using a Bayesian fusion approach to get the final segmentation of the lung region. Experiments on eleven real data sets based on Dice Similarity Coefficient (DSC) metric confirms the high accuracy of the proposed approach. The execution time results show that our algorithm takes about three seconds which is about 103 times faster when compared to a naive single threaded implementation on CPU.
AB - A new framework implemented on NVIDIA Graphics Processing Units (GPU) using CUDA for the precise segmentation of lung tissues from Computed Tomography (CT) is proposed. The CT images, Gaussian Scale Space (GSS) data generation using Gaussian Kernels (GKs), and desired maps of regions (lung and the other chest tissues) are described by a joint Markov-Gibbs Random Field (MGRF) model of independent image signals and interdependent region labels implemented on GPU. The initial segmentation from the original and the generated GSS CT images is based on the Linear Combination of Discrete Gaussian (LCDG) models; The initial segmentation is obtained from the original and the generated GSS CT images; then they are iteratively refined using a parallel MGRF model implemented on GPU with analytically estimated potentials. Finally, these initial segmentations are fused together using a Bayesian fusion approach to get the final segmentation of the lung region. Experiments on eleven real data sets based on Dice Similarity Coefficient (DSC) metric confirms the high accuracy of the proposed approach. The execution time results show that our algorithm takes about three seconds which is about 103 times faster when compared to a naive single threaded implementation on CPU.
UR - http://www.scopus.com/inward/record.url?scp=84870045425&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84870045425&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-35428-1_11
DO - 10.1007/978-3-642-35428-1_11
M3 - Conference contribution
AN - SCOPUS:84870045425
SN - 9783642354274
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 86
EP - 93
BT - Machine Learning in Medical Imaging - Third International Workshop, MLMI 2012, Held in Conjunction with MICCAI 2012, Revised Selected Papers
T2 - 3rd International Workshop on Machine Learning in Medical Imaging, MLMI 2012, Held in conjunction with the 15th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2012
Y2 - 1 October 2012 through 1 October 2012
ER -