Limits...
Segmentation of Pulmonary Vascular Trees from Thoracic 3D CT Images.

Shikata H, McLennan G, Hoffman EA, Sonka M - Int J Biomed Imaging (2009)

Bottom Line: A quantitative validation was performed with more than 1000 manually identified points selected from inside the vessel segments to assess true positives (TPs) and 1000 points randomly placed outside of the vessels to evaluate false positives (FPs) in each case.On average, for both the high and low volume lung images, 99% of the points was properly marked as vessel and 1% of the points were assessed as FPs.Our hybrid segmentation algorithm provides a highly reliable method of segmenting the combined pulmonary venous and arterial trees which in turn will serve as a critical starting point for further quantitative analysis tasks and aid in our overall goal of establishing a normative atlas of the human lung.

View Article: PubMed Central - PubMed

Affiliation: Iowa Institute for Biomedical Imaging, The University of Iowa, Iowa City, IA 52242, USA.

ABSTRACT
This paper describes an algorithm for extracting pulmonary vascular trees (arteries plus veins) from three-dimensional (3D) thoracic computed tomographic (CT) images. The algorithm integrates tube enhancement filter and traversal approaches which are based on eigenvalues and eigenvectors of a Hessian matrix to extract thin peripheral segments as well as thick vessels close to the lung hilum. The resultant algorithm was applied to a simulation data set and 44 scans from 22 human subjects imaged via multidetector-row CT (MDCT) during breath holds at 85% and 20% of their vital capacity. A quantitative validation was performed with more than 1000 manually identified points selected from inside the vessel segments to assess true positives (TPs) and 1000 points randomly placed outside of the vessels to evaluate false positives (FPs) in each case. On average, for both the high and low volume lung images, 99% of the points was properly marked as vessel and 1% of the points were assessed as FPs. Our hybrid segmentation algorithm provides a highly reliable method of segmenting the combined pulmonary venous and arterial trees which in turn will serve as a critical starting point for further quantitative analysis tasks and aid in our overall goal of establishing a normative atlas of the human lung.

No MeSH data available.


Related in: MedlinePlus

A trajectory of vessel segments and value changes of the eigenvalues along the trajectory. Left-most white arrow in (a) shows the starting point of the trajectory that advances left to right. The trajectory is calculated with fixed σf = 1. λ3 oscillates as the front advances and takes negative values close to the junctions indicated by dark gray regions on the trajectory. λ3 takes large values at location (1) and (2) making absolute value of the ratio λ3/λ2 larger. At locations (3) and (4), the ratio λ2/λ1 becomes small.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2801012&req=5

fig2: A trajectory of vessel segments and value changes of the eigenvalues along the trajectory. Left-most white arrow in (a) shows the starting point of the trajectory that advances left to right. The trajectory is calculated with fixed σf = 1. λ3 oscillates as the front advances and takes negative values close to the junctions indicated by dark gray regions on the trajectory. λ3 takes large values at location (1) and (2) making absolute value of the ratio λ3/λ2 larger. At locations (3) and (4), the ratio λ2/λ1 becomes small.

Mentions: These criteria capture the characteristics of the model so as to differentiate tube structures from sheet and blob shaped structures when a point is within a vessel segment and close to its center. However, pulmonary vascular trees include many junctions, and criteria 2 and 3 above are not always satisfied around the branchpoints. λ1 and λ2 do not always satisfy criterion 3 above, especially in the thin segments and around the junctions. In addition, λ3 sometimes takes large positive values even when a point is within a straight segment, which implies criterion 2 is not always met. Figure 2(a) shows a trajectory by a white tube starting from a point demarcated by the left most white arrow. The tracking front advances left to right in the figure. The gray parts on the trajectory indicate the points where λ3 takes on negative values. These approximate to the junction locations. Figure 2(b) shows changes of the three eigenvalues of the Hessian matrix as the front point of the trajectory advances. λ1 and λ2 always take on large negative values compared to λ3. Intensity gradually changes towards junctions when the point is in a straight segment and it may cause λ3 to take on a positive value. When the point is around a junction, intensity decreases rapidly towards the segmental direction and it causes λ3 to take on a negative value. For these reasons, λ3 fluctuates along the trajectory. Also, λ2 oscillates with its phase opposite to λ3 since λ2 takes relatively smaller values when the point is around junctions. Based on this observation, the filter output function F(x) is defined as follows:


Segmentation of Pulmonary Vascular Trees from Thoracic 3D CT Images.

Shikata H, McLennan G, Hoffman EA, Sonka M - Int J Biomed Imaging (2009)

A trajectory of vessel segments and value changes of the eigenvalues along the trajectory. Left-most white arrow in (a) shows the starting point of the trajectory that advances left to right. The trajectory is calculated with fixed σf = 1. λ3 oscillates as the front advances and takes negative values close to the junctions indicated by dark gray regions on the trajectory. λ3 takes large values at location (1) and (2) making absolute value of the ratio λ3/λ2 larger. At locations (3) and (4), the ratio λ2/λ1 becomes small.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2801012&req=5

fig2: A trajectory of vessel segments and value changes of the eigenvalues along the trajectory. Left-most white arrow in (a) shows the starting point of the trajectory that advances left to right. The trajectory is calculated with fixed σf = 1. λ3 oscillates as the front advances and takes negative values close to the junctions indicated by dark gray regions on the trajectory. λ3 takes large values at location (1) and (2) making absolute value of the ratio λ3/λ2 larger. At locations (3) and (4), the ratio λ2/λ1 becomes small.
Mentions: These criteria capture the characteristics of the model so as to differentiate tube structures from sheet and blob shaped structures when a point is within a vessel segment and close to its center. However, pulmonary vascular trees include many junctions, and criteria 2 and 3 above are not always satisfied around the branchpoints. λ1 and λ2 do not always satisfy criterion 3 above, especially in the thin segments and around the junctions. In addition, λ3 sometimes takes large positive values even when a point is within a straight segment, which implies criterion 2 is not always met. Figure 2(a) shows a trajectory by a white tube starting from a point demarcated by the left most white arrow. The tracking front advances left to right in the figure. The gray parts on the trajectory indicate the points where λ3 takes on negative values. These approximate to the junction locations. Figure 2(b) shows changes of the three eigenvalues of the Hessian matrix as the front point of the trajectory advances. λ1 and λ2 always take on large negative values compared to λ3. Intensity gradually changes towards junctions when the point is in a straight segment and it may cause λ3 to take on a positive value. When the point is around a junction, intensity decreases rapidly towards the segmental direction and it causes λ3 to take on a negative value. For these reasons, λ3 fluctuates along the trajectory. Also, λ2 oscillates with its phase opposite to λ3 since λ2 takes relatively smaller values when the point is around junctions. Based on this observation, the filter output function F(x) is defined as follows:

Bottom Line: A quantitative validation was performed with more than 1000 manually identified points selected from inside the vessel segments to assess true positives (TPs) and 1000 points randomly placed outside of the vessels to evaluate false positives (FPs) in each case.On average, for both the high and low volume lung images, 99% of the points was properly marked as vessel and 1% of the points were assessed as FPs.Our hybrid segmentation algorithm provides a highly reliable method of segmenting the combined pulmonary venous and arterial trees which in turn will serve as a critical starting point for further quantitative analysis tasks and aid in our overall goal of establishing a normative atlas of the human lung.

View Article: PubMed Central - PubMed

Affiliation: Iowa Institute for Biomedical Imaging, The University of Iowa, Iowa City, IA 52242, USA.

ABSTRACT
This paper describes an algorithm for extracting pulmonary vascular trees (arteries plus veins) from three-dimensional (3D) thoracic computed tomographic (CT) images. The algorithm integrates tube enhancement filter and traversal approaches which are based on eigenvalues and eigenvectors of a Hessian matrix to extract thin peripheral segments as well as thick vessels close to the lung hilum. The resultant algorithm was applied to a simulation data set and 44 scans from 22 human subjects imaged via multidetector-row CT (MDCT) during breath holds at 85% and 20% of their vital capacity. A quantitative validation was performed with more than 1000 manually identified points selected from inside the vessel segments to assess true positives (TPs) and 1000 points randomly placed outside of the vessels to evaluate false positives (FPs) in each case. On average, for both the high and low volume lung images, 99% of the points was properly marked as vessel and 1% of the points were assessed as FPs. Our hybrid segmentation algorithm provides a highly reliable method of segmenting the combined pulmonary venous and arterial trees which in turn will serve as a critical starting point for further quantitative analysis tasks and aid in our overall goal of establishing a normative atlas of the human lung.

No MeSH data available.


Related in: MedlinePlus