I found a strange behavior about scattered images when calculating projection images of a water tube phantom for CBCT reconstruction. The intensity of the scattered images slightly differs depending on the projection angle, even if the phantom is rotationally symmetric. Especially, the intensity of the scattered images are slightly higher only when the projection angle is around 0 deg.
I checked primary images by substracting scattered images from total images, the primary images had no angle dependence.
This behavior only occurs when I rotate the water tube phantom. If I fix the water tube phantom and rotate the detector and x-ray source, the scattered image have no angle dependence.
I attached examples of total and scattered images of 0, 90, and 180 degree. The Window/Level are the same for all images. (I simulated half-scan mode, so the projection images are shifted.)
Is it possible to fix this behavior?