How to blur private details
Each image keeps its own reviewed regions and unresolved face choices while you continue through the visible queue in order. Optional detection presents face boxes for review. Manual marking is the authority for faces, plates and objects, and detection can miss faces.
Steps
- 1Continue to the next unfinished image
Follow the visible queue in order; unresolved face choices stay with their file if you switch or reorder the batch.
- 2Select suggestions or mark manually
Review optional face boxes and select only useful candidates, then draw, move or resize missing faces, plates, addresses, screens or objects.
- 3Confirm, apply and re-open
Hide editing guides, inspect and confirm every file, choose strong blur or pixelation, then export and open each downloaded file at full size.
Key decisions
Why is manual review required?
Face detection can be unavailable, miss visible people or propose the wrong boxes. It does not suggest plates, documents or other sensitive objects, so inspect the whole image and add manual regions.
Should I use blur or pixelation?
Choose the treatment that removes recognizable structure at the final display size. A subtle cosmetic effect is not sufficient for sensitive details.
Before you download
Know the limits
- Regions are never assumed to match across different images; every source must be reviewed separately.
- A detected box is only a candidate until you select it; selection does not remove the need to resize, remove or supplement the resulting region.
- A rectangle that is too small can leave identifying edges, text structure or nearby context visible.
Verify the copy
Common questions
Do detected face boxes become editable regions automatically?
No. Detection presents candidates for review. Only the boxes you select are added as editable face regions.
Are detected face boxes guaranteed to be complete?
No. Detection can miss faces or propose the wrong boxes, so add missing regions and inspect the whole image.
Can detection suggest license plates or objects?
No. License plates, addresses, screens and other objects still require manual regions.
Can one set of regions be copied across the batch?
No. Each source retains separate regions because the sensitive content and pixel geometry differ.
Does guided review detect or protect everything automatically?
No. It only moves to the next unfinished file. Suggestions can miss or misplace faces, and plates, addresses, screens and objects still require manual regions and whole-image confirmation.