How to find & delete duplicate photos

Duplicate and near-duplicate images quietly eat gigabytes — burst shots, re-saved edits, downloads grabbed twice. Here's how to clear them out on your own machine, without handing your library to a cloud service.

The three kinds of "duplicate"

Not all duplicates are identical files, which is why a simple filename or file-size check misses most of them. There are really three categories:

  • Exact duplicates — byte-for-byte identical files (the same photo copied into two folders).
  • Near-duplicates — the same image re-saved, resized, re-compressed, or lightly edited, so the bytes differ but the picture is the same.
  • Visually similar — different shots of the same scene, or images with very similar composition and color.

How PicSy detects each one

PicSy uses three complementary methods so you catch all three categories:

  • Content hashing (MD5) finds exact duplicates instantly, regardless of filename or location.
  • Perceptual hashing (pHash) compares the visual structure of images, so resized and re-saved copies group together even though their bytes differ.
  • Color-histogram matching surfaces visually similar images by comparing their color distribution.

Each method has an adjustable similarity threshold, so you can be strict (only true copies) or loose (anything that looks alike).

Step by step

  • 1. Add your folders. Point PicSy at the directories you want to clean; it builds a fast local index.
  • 2. Open Duplicates. Pick the method(s) — exact, perceptual, or histogram — and a threshold.
  • 3. Review the groups. Duplicates are grouped together with previews, sizes, and paths so you can see exactly what's matched.
  • 4. Remove what you don't need. Keep the best copy and delete the rest. Nothing is removed without your say-so.

Do it without uploading a thing

Cloud "cleanup" tools require sending your photos to someone else's servers. PicSy computes all hashes locally, so duplicate detection happens entirely on your device — fast, and private. See the full feature list or download PicSy to try it.

FAQ

Will it find resized or edited copies, not just identical files?

Yes — perceptual hashing and color-histogram matching catch resized, re-saved, and lightly edited copies, not only byte-for-byte duplicates.

Does it delete duplicates automatically?

No. It groups them for review and you choose what to remove — nothing is deleted without confirmation.

Does finding duplicates upload my photos?

No. Detection runs on your device using local hashes; your images never leave your computer.

Clean up your library today.

Free to download. Free forever with local models.