Limits...
An Example-Based Multi-Atlas Approach to Automatic Labeling of White Matter Tracts.

Yoo SW, Guevara P, Jeong Y, Yoo K, Shin JS, Mangin JF, Seong JK - PLoS ONE (2015)

Bottom Line: To further enhance the computational efficiency, we implemented our approach on the graphics processing unit (GPU).Through nested cross-validation we demonstrated that our approach yielded high classification performance.The average sensitivities for bundles in the left and right hemispheres were 89.5% and 91.0%, respectively, and their average false discovery rates were 14.9% and 14.2%, respectively.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Engineering, Korea University, Seoul, Republic of Korea; Department of Computer Science, KAIST, Daejeon, Republic of Korea.

ABSTRACT
We present an example-based multi-atlas approach for classifying white matter (WM) tracts into anatomic bundles. Our approach exploits expert-provided example data to automatically classify the WM tracts of a subject. Multiple atlases are constructed to model the example data from multiple subjects in order to reflect the individual variability of bundle shapes and trajectories over subjects. For each example subject, an atlas is maintained to allow the example data of a subject to be added or deleted flexibly. A voting scheme is proposed to facilitate the multi-atlas exploitation of example data. For conceptual simplicity, we adopt the same metrics in both example data construction and WM tract labeling. Due to the huge number of WM tracts in a subject, it is time-consuming to label each WM tract individually. Thus, the WM tracts are grouped according to their shape similarity, and WM tracts within each group are labeled simultaneously. To further enhance the computational efficiency, we implemented our approach on the graphics processing unit (GPU). Through nested cross-validation we demonstrated that our approach yielded high classification performance. The average sensitivities for bundles in the left and right hemispheres were 89.5% and 91.0%, respectively, and their average false discovery rates were 14.9% and 14.2%, respectively.

No MeSH data available.


An anatomic bundle consists of tracts with various shapes.Cingulum bundle (a) contains many tracts with different shapes as shown in (b).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4520495&req=5

pone.0133337.g002: An anatomic bundle consists of tracts with various shapes.Cingulum bundle (a) contains many tracts with different shapes as shown in (b).

Mentions: As shown in Fig 2, a bundle consists of tracts with various shapes. For convenience, tracts within a bundle are classified into groups according to their shape similarity, and thus a bundle is composed of tract groups. For representation simplicity and also labeling efficacy, each group is modeled with a multivariate Gaussian distribution. A mean vector of tracts within a group is calculated as follows:m=1N∑i=1Nfi(3)where N is the number of tracts in the group. The covariance matrix can be estimated with a well-known empirical unbiased covariance matrix.S=[sij]=1N-1∑k=1N(fk-m)(fk-m)T(4)However, the covariance matrix S becomes singular or near-singular if the number of tracts in the group is less than 3n or many tracts are nearly parallel to each other. To overcome this problem, we adopt a shrinkage estimation method [47] to find a substitute matrix S* as follows:S*=[sij*]={siiifi=jrij*siisjjifi≠j(5)andrij*={1ifi=jrijmin(1,max(0,1-λ*^))ifi≠jwithλ*^=∑i≠jVar^(rij)∑i≠jrij2where sii and rij are empirical unbiased variance and correlation respectively, and is an empirical unbiased variance of rij. The new matrix S* is guaranteed to have a full rank regardless of sample data. The similarity between a tract fi and a group Bj that is modeled with the multivariate Gaussian distribution can be calculated using the Mahalanobis distance dM:dM(fi,Bj)=(fi-mj)T(Sj*)-1(fi-mj)(6)where mj and are mean vector and covariance matrix of Bj respectively.


An Example-Based Multi-Atlas Approach to Automatic Labeling of White Matter Tracts.

Yoo SW, Guevara P, Jeong Y, Yoo K, Shin JS, Mangin JF, Seong JK - PLoS ONE (2015)

An anatomic bundle consists of tracts with various shapes.Cingulum bundle (a) contains many tracts with different shapes as shown in (b).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4520495&req=5

pone.0133337.g002: An anatomic bundle consists of tracts with various shapes.Cingulum bundle (a) contains many tracts with different shapes as shown in (b).
Mentions: As shown in Fig 2, a bundle consists of tracts with various shapes. For convenience, tracts within a bundle are classified into groups according to their shape similarity, and thus a bundle is composed of tract groups. For representation simplicity and also labeling efficacy, each group is modeled with a multivariate Gaussian distribution. A mean vector of tracts within a group is calculated as follows:m=1N∑i=1Nfi(3)where N is the number of tracts in the group. The covariance matrix can be estimated with a well-known empirical unbiased covariance matrix.S=[sij]=1N-1∑k=1N(fk-m)(fk-m)T(4)However, the covariance matrix S becomes singular or near-singular if the number of tracts in the group is less than 3n or many tracts are nearly parallel to each other. To overcome this problem, we adopt a shrinkage estimation method [47] to find a substitute matrix S* as follows:S*=[sij*]={siiifi=jrij*siisjjifi≠j(5)andrij*={1ifi=jrijmin(1,max(0,1-λ*^))ifi≠jwithλ*^=∑i≠jVar^(rij)∑i≠jrij2where sii and rij are empirical unbiased variance and correlation respectively, and is an empirical unbiased variance of rij. The new matrix S* is guaranteed to have a full rank regardless of sample data. The similarity between a tract fi and a group Bj that is modeled with the multivariate Gaussian distribution can be calculated using the Mahalanobis distance dM:dM(fi,Bj)=(fi-mj)T(Sj*)-1(fi-mj)(6)where mj and are mean vector and covariance matrix of Bj respectively.

Bottom Line: To further enhance the computational efficiency, we implemented our approach on the graphics processing unit (GPU).Through nested cross-validation we demonstrated that our approach yielded high classification performance.The average sensitivities for bundles in the left and right hemispheres were 89.5% and 91.0%, respectively, and their average false discovery rates were 14.9% and 14.2%, respectively.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Engineering, Korea University, Seoul, Republic of Korea; Department of Computer Science, KAIST, Daejeon, Republic of Korea.

ABSTRACT
We present an example-based multi-atlas approach for classifying white matter (WM) tracts into anatomic bundles. Our approach exploits expert-provided example data to automatically classify the WM tracts of a subject. Multiple atlases are constructed to model the example data from multiple subjects in order to reflect the individual variability of bundle shapes and trajectories over subjects. For each example subject, an atlas is maintained to allow the example data of a subject to be added or deleted flexibly. A voting scheme is proposed to facilitate the multi-atlas exploitation of example data. For conceptual simplicity, we adopt the same metrics in both example data construction and WM tract labeling. Due to the huge number of WM tracts in a subject, it is time-consuming to label each WM tract individually. Thus, the WM tracts are grouped according to their shape similarity, and WM tracts within each group are labeled simultaneously. To further enhance the computational efficiency, we implemented our approach on the graphics processing unit (GPU). Through nested cross-validation we demonstrated that our approach yielded high classification performance. The average sensitivities for bundles in the left and right hemispheres were 89.5% and 91.0%, respectively, and their average false discovery rates were 14.9% and 14.2%, respectively.

No MeSH data available.