Find image source and duplicates on your computer

The program helps find similar photos, for example, one original unedited photo on disks with Windows, Mac OS and Linux (example). Alternatively you can use Duplicate Finder to find multiple groups of similar images in a directory (image clustering).

Perceptual brightness similarity is used, e.g. images after minor color/spacial changes, resized pictures, and exact copies.

How to use

Image types supported

JPG, PNG, BMP, WEBP, GIF (static and animated) of maximum size 20MB per image. Transparency is ignored in image comparison.

Algorithm

Read how the image comparison algorithm works and check out image similarity in Go (Github) for your own offline file processing, company or project. The web implementation contains some improvements not described in the linked document above.

Precision

Parameters used in the demo web version are set to find reasonably high number of near-similar images with acceptable speed and low false positives. For Clustering one image appears in one cluster only.

Images are compared by perceptual similarity in terms of color and brightness across the whole image, without specifically taking in account such transformations as rotations and crops. Partial occlusions, e.g. watermarks, impact perceptual similarity. If the perceptual changes are small, the pictures are considered similar.