Image Redactor - Support Multiframe Image Redaction

[joelbyler]

Is your feature request related to a problem? Please describe.

I'm attempting to use presidio image redaction in cases where a dicom instance contains multiframe pixel data, an example of this would be a XA modality instance. There are limitations with this library which prevent image redaction on multiframe instances, rather only one frame is used for PHI detection and the bboxes may not be applied properly or to all frames. While it may not be common for multiframe instances to include PHI, it is possible for some PHI to be "burned in" to the pixel data on a multiframe instance.

Describe the solution you'd like
A clear and concise description of what you want to happen.

Rather than ignoring the extra frames on a multiframe instance, each frame that is included in a multiframe instance should be scanned for PHI and have that PHI redacted.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

I believe there are no workarounds in this case other than to not use this library for multiframe dicom instances

Additional context
Add any other context or screenshots about the feature request here.

I'm including test files which were computer generated with fake pixel data and synthetic PHI

multiframe_test_images.zip

Additionally, this was created as a spin off issue from #1735

[vatsalm1611]

Hi — I'd like to work on this issue.

Plan:

• Investigate current multiframe handling in the image redactor.
• Add support to iterate over all frames in a multiframe DICOM, run detection on each frame, and apply redaction per-frame.
• Add unit tests for at least one multiframe DICOM scenario (mocked or small test files).
  I will create a small, focused PR with tests and implementation. Please let me know if maintainers prefer a different approach.