DupDetector: find duplicate and near-duplicate images

Built for large photo collections: finds identical and near-identical files, even when the copy was resized, brightened or sharpened.

DupDetectorScan: 4 pairs
PairFilesSimilarityDifference
IMG_2041.jpg
IMG_2041 (copy).jpg
100%Exact copy
beach_01.tif
beach_01_web.jpg
97%Downscaled
old_town.png
old_town_edit.png
93%8% brighter
cat_window.bmp
cat_window_sharp.jpg
91%Sharpened
3.3current version
1.1 MBinstaller size
10image formats
3 MBon disk after install
Near-duplicates are duplicates too

Even commercial stock libraries contain identical shots that differ only in sharpness, brightness or a slight color shift. DupDetector catches those pairs as well.

Original lake photoBrightened copy OriginalCopy
Pair similarity

You set the threshold: strict for exact copies, looser for bursts.

97%

A safe way to clean up

Report first, decisions second. The program shows you pairs and groups; deleting or moving files is up to you.

  • ScanPick folders or an entire drive and DupDetector compares the file contents.
  • Review pairsOriginal and copy side by side, with similarity score and dimensions.
  • Quarantine, don't deleteMove extras to a separate folder first, just to be safe.
Content-based comparison

File names don't matter: IMG_2041 and “Sea final 2” may well be the same photo.

IMG_2041.jpg ≈ Sea final 2.jpg

Where duplicates come from

Nobody saves the same photo four times on purpose. These are the usual suspects.

Memory cards imported twice

The same card copied to a laptop, then again after an OS reinstall.

Sync and backup tools

Cloud folders and backup apps that restore whole trees next to the originals.

Edited exports

A brightened or resized copy saved beside the source under a new name.

Messengers and downloads

The same picture received in three chats, each recompressed a little differently.

Choosing a similarity threshold

Starting points for a typical family archive. Stock libraries and bursts may need a stricter setting.

SimilarityWhat it usually isAction
99-100%Exact copiesSafe to quarantine in bulk
93-98%Resized or recompressedKeep the largest file
85-92%Edited or burst framesReview pair by pair
below 85%Similar scenesUsually not duplicates

A cleanup order that keeps originals safe

GatherCopy every source into one inbox folder first, so the comparison sees everything.
Scan strict, then looseClear exact copies at a high threshold before looking at near-duplicates.
Pick the keeperPrefer the largest dimensions, the earliest date and the original file name.
Wait, then deleteLeave the quarantine folder for a few weeks before emptying it.

Frequently asked questions

How does DupDetector find near-duplicates?

It compares images by their content, so a copy with different brightness, sharpness or size is still paired with the original.

Does it delete files on its own?

No. DupDetector shows you the pairs and groups it found; you decide what happens to each file.

Which formats are supported?

JPG, GIF, BMP, PNG, TIF, PCX, TGA, WMF, EMF and PSP.

What are the system requirements?

Windows 95 or later, 32 MB of RAM and 3 MB of free disk space.

The rest of the Prismatic toolkit

Three tools cover a photo from start to finish: processing, archive cleanup and mosaic building.