1887
banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
3D automatic anatomy segmentation based on iterative graph-cut-ASM
Rent:
Rent this article for
USD
10.1118/1.3602070
/content/aapm/journal/medphys/38/8/10.1118/1.3602070
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/38/8/10.1118/1.3602070

Figures

Image of FIG. 1.
FIG. 1.

The flowchart of the proposed AAS system.

Image of FIG. 2.
FIG. 2.

Different slices of intensity weighted b-scale scenes extracted from a CT image (female subject, abdominal region) are shown in the first row. Second–fifth rows show corresponding thresholded intensity weighted b-scale scenes for increasing thresholds. The last row denotes some truly segmented objects of the abdominal region.

Image of FIG. 3.
FIG. 3.

Example textured shapes. Geometric and appearance based centroids are shown in diamond and square respectively. Region based centroids are obtained by weighting shape points’ coordinates with the corresponding intensity values.

Image of FIG. 4.
FIG. 4.

The shape and intensity structure systems, PAo and PAb , are shown in the spherical coordinate system with their Euler angles drawn in similar colors. Any orientation difference between the PA systems requires the computation of another orthonormal rotation denoted Rob , which rotates the shape structure system into alignment with the intensity structure system on the sphere.

Image of FIG. 5.
FIG. 5.

(a) A CT slice of the abdominal region with selected objects (skin, liver, spleen, and left and right kidneys) is shown on the left. Annotated landmarks for the selected objects are shown on the right. (b) An MRI slice of the foot with selected objects (skin, navicular, calcaneus, tibia, talus, and cuboid) is shown on the left. Annotated landmarks for the selected objects are shown on the right.

Image of FIG. 6.
FIG. 6.

Mean shape is generated using 3D-ASM for multiple objects of the abdominal region. (a) Mean shapes of liver, spleen, right and left kidneys. (c) Mean shapes of calcaneus, talus, tibia, navicular, and cuboid. (b) and (d) Mean shapes of skin boundaries for the objects presented in (a) and (c), respectively.

Image of FIG. 7.
FIG. 7.

(First column) the model assembly (MA) is overlaid with the organs/objects of one subject prior to recognition. (Second column) positioned MA for the subject is shown after recognition.

Image of FIG. 8.
FIG. 8.

The experimental results for multiorgan segmentation are shown in three different anatomical levels for CT abdominal dataset. The first column shows original images slices; the second column indicates the recognized organs; and the third column shows the delineation results yielded by the proposed IGCASM. The contours in third column shows manually delineated organ boundaries. All of the images have been cropped for the best view and original image size is (512 × 512).

Image of FIG. 9.
FIG. 9.

The experimental results for multiorgan segmentation are shown in three different anatomical levels for foot MRI dataset. The first column shows original images slices; the second column indicates the recognized organs; and the third column shows the delineation results yielded by the proposed IGCASM. The contours in third column shows manually delineated bone boundaries. All of the images have been cropped for the best view and original image size is (512 × 512).

Image of FIG. 10.
FIG. 10.

Three different views of delineation results for three examples: CT abdominal organs by IGCASM.

Image of FIG. 11.
FIG. 11.

Three different views of delineation results for two examples: MRI foot bones by IGCASM.

Tables

Generic image for table
TABLE I.

Different numbers of landmarks for different objects are listed.

Generic image for table
TABLE II.

Quantitative measure of the proposed recognition method.

Generic image for table
TABLE III.

Mean and standard deviation of delineation results as TPVF, FPVF for 3D ASM, and IGCASM on CT abdominal and foot MRI data.

Loading

Article metrics loading...

/content/aapm/journal/medphys/38/8/10.1118/1.3602070
2011-07-25
2014-04-20
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: 3D automatic anatomy segmentation based on iterative graph-cut-ASM
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/38/8/10.1118/1.3602070
10.1118/1.3602070
SEARCH_EXPAND_ITEM