Generation of metal artifacts in dental CT

[2177361170]

Hello,

I noticed that most of the images in the AAPM Challenge dataset are of lungs, abdomen, and head, with very few dental CT images.

Therefore, I would like to use your tool to simulate metal artifacts on dental CT images. How can I proceed?

[MingyeWu]

Hi there,
For your needs, usually the steps are:

1. Collect the dental images that DO NOT have metal and convert them into voxelized phantoms.
2. Add metal objects into the voxelized phantoms.
3. Simulate the images with the metal.

The reason for not having metal in the original images: if they have metal, the metal artifacts will make it almost impossible to generate the phantoms because the phantoms are supposed to be the ground truth.

[2177361170]

Yes, I am conducting the experiment according to the steps you provided. I would like to use the data and methods you provided to generate metal artifacts on the head imaging data first, and then perform oral imaging data.

Firstly, I used the DICOM_to-voxelized_phantom.py method provided by the phantoms-voxelized in the xcist repository and the default dataset to generate a voxelized phantom of the head, as shown in the following figure.

Secondly, I didn't understand what you meant by 'Add metal objects into the voxelized phantoms'. Is this to determine the size and position of the metal objects?

Thirdly, 'Simulate the images with the metal.' should be done by running experientiant.analyticArtifactPhantom. py, and I also successfully ran it using the default configuration, obtaining 6 independent experimental results. But when I tried to run other image files, I found that phantom.filename was a. ppm file type, but there was no PPM file in the voxelized phantoms I generated. I tried to change it to a JSON file but found it couldn't run properly. So how can I obtain this PPM file? Was it generated in the second step? And The PPM file does not seem to be CT image data.

[MingyeWu]

I cannot see your image, but I guess it might be the known issue:

The Root Cause:
We identified that the new streak artifact was caused by the low-signal correction, which was introduced after the evaluation study and is enabled by default.
In this simulation, there are metal rods that cause extremely high attenuation. However, since this is an ideal case with no noise and sufficient precision, the low-signal correction is unnecessary. In fact, because the correction replaces low-signal pixels with the average of neighboring pixels, it introduces error and causes streak artifacts.

The Solution:
We added a cfg flag (cfg.physics.disable_low_signal_correction) to disable the low-signal correction when needed. We updated the following code:
pyfiles/PrepView.py
The evaluation sample script (experiment_analyticArtifactPhantom.py)

So, with the latest code, just set cfg.physics.disable_low_signal_correction=1