Recently I upgrade the Pylinac from Ver 2.2.6 to Ver 3.0.0.
With 2.2.6 version :
I have a problem with ROIs angle for phantom positioned on the treatment couch, i.e. 90-degree gantry angle. please see the “RI.G90.DCM” image. Phantom positioned on 100cm SSD.
for Ver 3.0.0:
I tried to repeat this test to analyze the same image as before, but this time the ROIs scale was very small in comparison with a phantom outline. I also passed the SSD argument in the “analyze ()” function but didn’t work.
furthermore, I tried to analyze the same phantom. this time the phantom is positioned on the EPID surface (RI.G0.dcm image) and the Pylinac Ver 3.0.0 Unable to find the phantom in the image and return an error.
Hi Br,
I’m sorry for the late reply–for some reason this slipped through the cracks. I was able to reproduce your results and indeed pylinac is having trouble with your G90 image. It seems that the edge detection is getting caught on the acrylic phantom holder. See the attached image of the top half of where the phantom and holder separate. This edge is being included in the result and is what is causing the ROI to not meet the conditions for finding the phantom. I tweaked a few things but nothing obvious fixed it. I will continue working on this but at least wanted to give you an update. Outside of fixing the issue, the other possible resolution is to use some sort of shim between the phantom jig and phantom itself. E.g. a folded-up piece of paper underneath the phantom to give some space. I realize that is far from ideal but wanted to give you some sort of possible workaround. Thanks for bringing this to my attention!
