UCL / PETPVC

Partial Volume Correction in PET
Apache License 2.0
53 stars 14 forks source link

Long runtime and incorrect output format for PETPVC rbv correction with 4D images #91

Open ninea-anasovi opened 1 year ago

ninea-anasovi commented 1 year ago

Description:

I am encountering two issues with PETPVC when applying region-based voxel (rbv) correction to 4D PET dynamic images. The PET image consists of 28 frames, and I have a corresponding 4D mask image with the same number of frames.

Long Runtime:

When I attempt to perform the rbv correction using PETPVC with 28 frames, the runtime exceeds 3 hours and the process does not complete. I have left my computer running overnight, but the correction does not finish. Incorrect Output Format: As an alternative, I tried applying the rbv correction with only 2 frames. Although the correction completed in a reasonable time frame, the resulting image was in a 3D format instead of the expected 4D output.

Steps to Reproduce:

Prepare a 4D PET dynamic image with 28 frames. Prepare a 4D mask image with 28 frames. Apply the rbv correction using PETPVC. Observe that the correction process takes longer than 3 hours without completing for 28 frames. Additionally, observe that when using 2 frames, the rbv correction finishes, but the output is in 3D format instead of 4D.

Additional Information:

I have confirmed that both the PET dynamic image and the mask image are in the correct 4D format required by PETPVC. I have reviewed the PETPVC documentation and did not find any specific instructions regarding the handling of 4D data. I am using the latest version of PETPVC (please specify the version number if known).

Expected Behavior:

The runtime for the rbv correction should be reasonable, even for 28 frames. The rbv correction should generate a 4D corrected PET image, preserving the temporal dimension of the input data. Please let me know if any further information or clarification is needed to address these issues. Thank you!