[Insight-users] GeodesicActiveContourShapePriorLevelSetImageFilter

[PY]

Hi all,


    I am attempting to experiment with the 
GeodesicActiveContourShapePriorLevelSetImageFilter class, and I have ran 
into some troubles.  I gave in one image as the shape prior image and 
ran the filter with these parameters

Usage: GeodesicActiveContourShapePriorLevelSetImageFilter.exe
inputImage: stickprostate.png (a prostate image)
outputImage: stickout.png
seed1x: 290
seed1y: 198
seed2x: 399
seed2y: 301
seed3x: 494
seed3y: 199
initialDistance: 5
sigma: 1
propagationScaling: 0.5
shapePriorScaling: 0.02
meanShapeImage: shape0.png
numberOfModes: 1
shapeModeFilePattern: shape%d.png
startX: 0
startY: 0

    The output I got with this was a pure black image.  My shape prior 
image was a simple prostate I drew by painting over the prostate in the 
original input file and saving it in another file.  My startX and startY 
parameters were (0,0) because I didn't feel any translations were 
necessary since I basically painted over the original image.  The seed 
points should be valid as well as I have used them before on the 
GeodesicActiveContourImageFilter, but the problem is that the contour 
"leaks" with that filter.  So does anyone have an idea why there is no 
segmentation whatsoever with the shape prior method I'm using? 

- PY