Our UR5e image set mixes two polished-cover finishes under the same scratch labels, and the bright line shifts when either cover is reseated; who should decide which physical examples belong in the comparison before I let those labels near the software?
Quality should identify the samples and their actual acceptance decisions, with the vision reviewer defining the comparison; are both finishes intended for this station?
Give each intended finish its own identified examples, because our recent comparison needed representative finishes as well as the difficult reject; one attractive supplier image cannot stand in for both.
Yes, incoming records identify the two retained sets; quality is examining them separately and the unidentified supplier image is out of the working reference set.
Keep the faint real defect in the lighting comparison when quality identifies it, because on ours reducing the stripe also made the edge scuff too faint to use.
They've identified a scratch on one finish and accepted the retained examples of the other; we still need a suitable reject for that second finish before calling the set representative.
Second-finish reject now retained and judged by quality; both finish groups are going into the presentation comparison, with the original unidentified photo kept only in the correspondence archive.