When PhotoSanitize identifies potential sensitive content, you decide which regions should be concealed. The Visual Privacy review is intended to help you make that decision before permanent changes are applied to the sanitized output.
The review is especially important with automatic detection, because identifying a face, text region, or barcode is not the same as deciding that it must be hidden.
Reviewing a photograph
- Open Visual Privacy for the analyzed image.
- Run Detect if automatic detectors are enabled and required.
- Examine the complete picture and the detection findings.
- Select the regions that should be redacted and deselect those that should remain visible.
- Adjust the regions where necessary and add manual redactions for anything that was missed.
- Choose Mark Reviewed when the required review is complete.
For example, a company photograph might include employees who have been approved for publication and a visitor whose identity should not be disclosed. Automatic detection may identify all their faces, but you can decide which should be concealed.
The same principle applies to text. A photograph may contain both ordinary signs and confidential names. Their presence in the detection list does not mean they all require the same treatment.
Examining missed information
Do not limit your review to the areas identified automatically. Look at the background, reflections, small documents, screens, windows, badges, and other details that could reveal information.
A photograph can also contain identifying clues that do not fit the available detector categories, such as a distinctive house, company vehicle, or recognizable workplace interior.
In such situations, use manual redactions where appropriate or select a different photograph for distribution.
Understanding Mark Reviewed
Mark Reviewed records that you have completed the required review of the image’s visual findings.
It does not mean that PhotoSanitize has independently determined whether the entire image is private, anonymous, or suitable for publication.
The confirmation applies to the current review state. If you subsequently change the relevant detector settings, selected findings, or redaction regions, the previous review may need to be repeated.
For collections containing several photographs, review is handled for each relevant image. Approving one image does not approve every file in the batch.
Maximum Privacy
The Maximum Privacy profile requires automatic detection of faces, license plates, visible text, QR codes, and barcodes, followed by human review.
Sanitization cannot proceed under this profile until the required review has been completed. If one of the included images still needs attention, PhotoSanitize directs you to the outstanding review rather than silently ignoring it.
This requirement also explains why Maximum Privacy is not available through the headless Enterprise command-line interface. Its workflow depends on an actual user reviewing the picture and redaction selections.
Checking the final result
After sanitization, open the resulting image rather than relying only on the preview. Redaction selections become permanent pixel changes in that output.
Check the boundaries of the concealed areas and review the rest of the picture for remaining information. If something was missed, return to the original, correct the selection, and sanitize again.
For organizational publication, it may be appropriate for someone other than the person who prepared the photograph to review the final output. PhotoSanitize provides the technical redaction and verification tools, while the decision to release the image remains part of the organization’s own workflow.