Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

System workflow.

More »

Fig 1 Expand

Fig 2.

Left: Sphere model for vessel. The foreground is computed inside of sphere s and the background is computed using the data in ΔI and outside of tube. Right: vessel in a 3-D voxel window without enforced noise.

More »

Fig 2 Expand

Fig 3.

A vessel in a 3-D voxel window with directional constraint.

More »

Fig 3 Expand

Fig 4.

Noisy image makes a vessel appears to be disconnected.

More »

Fig 4 Expand

Fig 5.

Discontinuity in LAD (left anterior descending) artery shown.

(e) is a magnified version of (d).

More »

Fig 5 Expand

Fig 6.

Active search for vessel candidate.

Type 1 and 2 active search methods are illustrated. Type 1 is used for disconnected vessel and type 2 is used for disconnected branches.

More »

Fig 6 Expand

Fig 7.

Example of tracked centerlines and actively searched region.

Actively searched regions are marked in blue circles. Note there are many false positive seed points (white and yellow spheres). They are automatically deleted later.

More »

Fig 7 Expand

Table 1.

Dataset used for Poisson parameter estimation.

Poisson pdf () trained on LM (left main) artery length.

More »

Table 1 Expand

Fig 8.

Planar vesselness (red) vs. conventional 3-D Frangi’s Vesselness (blue) along the vessel centerline.

Both have high values along the tracked vessel. And they are much correlated at proximal and mid-part of coronary artery. When there is no vessel, both values are near zero.

More »

Fig 8 Expand

Fig 9.

Exhaustive search for initial direction of vessel using cylinder model.

More »

Fig 9 Expand

Table 2.

Test result using 8 Rotterdam Training dataset.

It shows detailed values for each vessel. (sc: score.).

More »

Table 2 Expand

Table 3.

Average overlap per Test dataset.

More »

Table 3 Expand

Table 4.

Standard deviation and p-values for all parameters OV, OF, OT (Table 3).

More »

Table 4 Expand

Table 5.

Average accuracy per Test dataset.

More »

Table 5 Expand

Table 6.

Summary for Test dataset.

More »

Table 6 Expand

Fig 10.

Examples of coronary artery structure found by the proposed method.

More »

Fig 10 Expand

Fig 11.

Example of centerline extracted (left) and coronary artery tree mesh reconstruction (right) found by the proposed method.

More »

Fig 11 Expand

Table 7.

Overlap Measure Scores for all the methods using Test dataset.

Also accuracy distance measure is shown in mm. Last column is marked with a circle for the similar experiment to us.

More »

Table 7 Expand