Limited angle CBCT reconstruction issue #154
Comments
|
To fix this issue insert the following command before you run any of the projectors The issue is that you are doing an offset scan, but offset scans only work if you have projections over 360 degrees or more, so when you specify less than 360 degrees, LEAP essentially turns offset scan to false and uses a smaller field of view. The command above turns off the field of view mask. |
|
Hi Kyle, Thank you very much for your help. It works. |
Sign up for free
to join this conversation on GitHub.
Al
4313
ready have an account?
Sign in to comment
I tried to follow the closed issue #40 for limited angle CBCT reconstruction. In #40, fan-beam geometry was used and I use conebeam geometry:
numRows = 768
numCols = 1024
numAngles = 360
angles=360.0
pixelSize = 0.388
device_1=th.device("cuda:0")
leapct.set_conebeam(numAngles, numRows, numCols, pixelSize, pixelSize, 0.5*(numRows-1), 0.5*(numCols-1)-412, leapct.setAngleArray(numAngles, angles), 1000, 1500)
leapct.set_offsetScan(True)
leapct.set_volume(256, 256, 87, voxelWidth=2*0.908, voxelHeight=2.0)
leapct.print_parameters()
g = leapct.allocateProjections()
f = leapct.allocateVolume()
f = leapct.load_volume('..\pro_001_cbct_crop_fit.npy').astype(np.float32)
f = f/30
f = np.ascontiguousarray(f)
leapct.project(g,f)
If numAngles and angles are 360, the projections are correct:
but if numAngles and angles are other than 360, the projections are very strange and incorrect:
Could you let me know if there are some setting incorrect and how can I fix this issue?
Thank you very much.
Ryan
The text was updated successfully, but these errors were encountered: