The proposed report would allow us to catch:

• Merged releases where both source and target already had cover art
• Redundant uploads made to releases which already had cover art

While doing some routine cleanup, I noticed this is more widespread than I would have thought (although I cannot produce any statistics).

There would likely be many false positives (e.g. box sets where each medium also has its own front/back covers), so this would be most helpful in conjunction with MBS-2919. Hopefully the amount of false positives would be limited by only checking front/back images and not other types of images.