Image Hash Comparator (Perceptual pHash, dHash and aHash)

Compare images with perceptual hashes.

Use the tool

Why this tool exists

Choose two or more images. The page computes a pHash, dHash or aHash for each from a tiny grayscale copy and ranks the pairs by Hamming distance, the number of differing bits out of 64.

Small distances mean similar pictures, so resized or re-compressed copies match. Crops, rotations and mirrored copies do not. The near-duplicate threshold is a rule of thumb, so review the matches by eye.

How to use it

  1. Choose at least two images.
  2. Pick the hash type.
  3. Set the near-duplicate threshold.
  4. Read the closest pairs.

Frequently asked questions

Which hash should I use?

pHash is the most robust to edits. dHash is fast and good for copies. aHash is the simplest.

What distance means a duplicate?

0 is the same hash. Up to about 5 to 10 bits is often a near copy. Always check.

Does it find rotated copies?

No. The hashes change when the picture turns.

Is anything uploaded?

No. The work happens in your browser and the page makes no network request with your input.