Home › Resources › Glossary › Duplicate detection
Duplicate detection identifies files that are identical or sufficiently similar to warrant review. In digital asset management (DAM), it helps teams spot repeated uploads and distinguish unnecessary copies from useful versions or formats. A match is a prompt to check the asset, not an instruction to delete it.
HOW IT WORKS
Repeated copies can split useful information across several records. One may hold the licence, another the campaign tags, and a third the approval decision. Finding a match is only the first step: the team must decide which record to retain and what information needs preserving.
An exact-file check compares file content using a digital fingerprint, often called a hash. Renaming a file does not change its content. Visual matching compares image appearance instead, so it can flag some changed files that an exact check would miss.
A revised design, translated document or channel-specific crop may be intentionally different. Version control records revisions; duplicate detection looks for possible repetition. Neither tells you whether a file is approved or licensed for your intended use.
IN PRACTICE
Asset Bank is a DAM platform for marketing, brand and creative teams. Its duplicate detection guidance describes two matching options:
The documented perceptual method does not match a cropped image to its uncropped original. Sensitivity affects missed matches and false positives. Ask Asset Bank support to confirm the configuration that applies to your library.
For stored files, administrators can use Admin > Reports > Asset > Duplicate Assets. See the report instructions. These features support review; they do not decide which licence, approval or version you should retain.
ILLUSTRATIVE EXAMPLE
A university communications team receives a photograph already held in its DAM. The incoming file has a different name, and the folder also contains a smaller web export and a square social crop.
The reviewer confirms that the renamed file is identical to the stored original. Before discarding it, they check whether the incoming record adds any credit or consent information. They retain the original record with its approved metadata.
The web export and social crop have distinct uses. The team checks their quality and intended channels, then decides whether to retain them as labelled variants or regenerate them from the original when needed. A visual match alone would not justify deleting either.
Result: the redundant copy is removed after review, while useful outputs and their usage context are preserved. This is an illustrative workflow, not a customer result or a claim that every variant will be detected.
Not when the check compares file content. Asset Bank’s data-matching option identifies identical files even if their filenames differ. A renamed copy is still a copy; a genuinely edited file may require a different comparison.
No. Matching depends on the method and its settings. Asset Bank documents perceptual matching for resized or recoloured images, but says it will not match a cropped image to its uncropped original. Review flagged results and do not assume an unflagged file is unique.
No. Compare the files and their records first. Check quality, intended use, rights, approval, metadata and existing links. Preserve useful variants and necessary information before removing a redundant copy through your agreed review process.
Explore how Asset Bank helps marketing teams manage, control and share digital assets.
KEEP EXPLORING
Sources checked: 1 November 2026.
