Our FR5 image folder contains two different faceplates under one sample name. Bright streaks shift between views as well. I am stopping threshold work until we reconstruct which physical examples these labels describe.
I once traced a disputed mark to a reused sample bag rather than an imaging fault. Keep the physical plates separate and compare their identifiable features with the original captures before assigning either scratch decision to a particular plate.
Leave those remaining views unattributed for now. Do the inspected parts have distinct quality decisions, or has the shared name confused that record too?
What changed between the views with the travelling streak: plate position, lamp position, or both? That detail may still be recoverable even where the plate identity is not
Carla, quality inspected the notched plate and accepted it. The other has no decision. Leah, the capture notes record reseating with the lamp fixed for the identifiable views.
Then the accepted plate can support a new labelled set, with its identity retained, while quality assesses the other plate independently; the old scratch labels still need withdrawing wherever they were copied.
Omar, it appears in the draft training pack. I have withdrawn that page. Both plates now have separate identifiers linked to rear and front photographs.
The draft page now says references under review. We have retained the original pack separately. The supplier has been told not to use its scratch examples.
And obtain at least one quality-confirmed rejected feature for the imaging comparison, because identifying an accepted plate does not show whether the proposed view reveals the defects you need to detect.
Quality owns the physical decisions; I own the linked image set. Their sample selection includes finish variation and rejected marks. The currently unassessed plate remains separate.
Is the reseating method representative of the proposed tool, or are you still moving the plates by hand? That affects what your comparison can tell the designer
Hand placement in a prototype nest. Its side clearance permits visible tilt. Tooling is reviewing the seating arrangement before we collect the replacement set.
Record the current nest revision with the existing captures so the subsequent comparison can distinguish a seating change from a different physical sample.