Negative values after simulation

[lhr522857918]

Hi

Thanks so much for the hard work.

I am encountered with this issue that after I simulate the projections with the scatter, the air area around the object has negative values. I am using the simulation code in the example while the simulation of noise-free version turns out to be successful. Have you any idea what could be the problem. I attach the code. ct1 = xc.CatSim("/home/lhrrr/Desktop/task1/json/1/para") # initialization

--------- Make changes to parameters (optional)

ct1.resultsName = "test_sc" ct1.physics.scatterCallback = "Scatter_ConvolutionModel" # scatter model ct1.physics.scatterKernelCallback = "" # using default ct1.physics.scatterScaleFactor = 80

ct1.protocol.viewsPerRotation = 32 ct1.protocol.viewCount = ct1.protocol.viewsPerRotation ct1.protocol.stopViewId = ct1.protocol.viewCount-1

ct1.scanner.detectorRowsPerMod = 512 ct1.scanner.detectorRowCount = ct1.scanner.detectorRowsPerMod

ct1.scanner.detectorColPerMod = 512 ct1.scanner.detectorColCount = ct1.scanner.detectorColPerMod

--------- Run simulation

ct1.run_all() # run the scans defined by protocol.scanTypes

scatted image Scatter free image from same recon image

[zhangjy-ge]

Hi,

It seems to me those negative pixels are due to too high scatter in your simulations, specifically the parameters of ct1.physics.scatterScaleFactor = 80 seems too large to me. If there are no real reasons, you may try to set it to 1 and see how it goes.

Best, Jiayong

[lhr522857918]

Hi

Thanks so much for your quick response! May I ask what the scatter factor means physically.

I tried factor = 1, some pixels still have negative values, what is a suggested nominal value to use for cone beam？

Thanks again. Best.

[zhangjy-ge]

The scatter scale factor means w in the following formula [https://doi.org/10.1088/1361-6560/ac9174]:

which is used to adjust SPR. A scale factor of 1 roughly corresponds to SPR of 15%, so you might change this factor linearly accordingly based on your specific SPR, but 80 is a too large number. Since we don't know your system and SPR, we cannot suggest a value.

Another thing you could try is to add scatter correction, by setting: physics.callback_pre_log = "Scatter_Correction" physics.scatterCorrectionScaleFactor = 1

It is normal to have some negative values due to scatter. As long as you have reasonable scale factor (and correction), those negative values should be very small or gone.

[lhr522857918]

Hi, I checked with scatter factor = 1 and scatter correction, some parts of the projection is still negative. I am wondering how you converted the detector signal(yi) into the projection(the prep in your simulation), since I checked that each element in the equation shall be all non-negative. I am guessing that the step converting yi to the output projection may cause negative values, could you please tell me how you did it.

Thanks again for your great effort.

Best.

[MingyeWu]

Hi, Thanks for testing this. The negative value in the p-value (after -log) is due to scatters cause some pixels to have even more signal than airscan, which happens around the edge of the object, unless there's a perfect scatter correction. I have updated PrepView.py to avoid this issue by simply zeroing all negative p-values.

Mingye

[lhr522857918]

Hi Thanks so much for your clarification and remarkable work!

Best