Scan a folder of pictures and find exact copies as well as near-duplicates that were resized, recompressed, lightly edited or cropped. Matching photos are grouped, the best copy is suggested to keep, and you can compare any pair side by side before cleaning up.
Last updated:
Scan a folder or drop photos hereJPG, PNG, WebP, GIF, BMP, AVIF · photos are read on your device and never uploaded
identical onlyloosely similar
Side-by-side review
Advertisement
Advertisement
How to use the Duplicate Photo Finder
Press Scan folder and pick the folder with your photos (subfolders are included), or drop a folder or photos onto the box. A small demo set is scanned on load to show how results look.
Wait for the scan: each photo is decoded once and reduced to a 64-bit fingerprint. Thousands of photos take a minute or two.
Set the sensitivity: around 4-6 bits finds the same photo in different sizes or qualities; 10 or more also finds burst shots and light edits.
Review the groups: the suggested keeper has a green border. Click any thumbnail to compare it side by side with the keeper, and untick Duplicate for photos you want to keep.
Export the list with Copy marked file list or Download CSV report, then delete those files in your file manager or photo app.
How perceptual hashing finds duplicates
A normal file hash such as MD5 or SHA-256 changes completely when a single byte changes, so it only finds exact copies. A perceptual hash describes what the picture looks like instead. This tool uses the difference hash (dHash):
The photo is scaled down to 72 × 72 pixels and converted to brightness (0.299 R + 0.587 G + 0.114 B).
Blocks of 8 × 8 pixels are averaged into a 9 × 9 grid, which throws away fine detail, noise and compression artifacts.
For each of the 64 cells in an 8 × 8 area, one bit records whether the cell is brighter than its right-hand neighbor (horizontal hash) and another whether it is brighter than the cell below (vertical hash).
Two photos are compared by counting the differing bits, the Hamming distance. The tool averages the horizontal and vertical distances, so the result is 0 to 64 bits.
Photos whose distance is at or below the sensitivity are joined into a group (if A matches B and B matches C, all three end up in one group). Within a group, files with the same size are also checked with SHA-256 to label byte-identical copies.
What the difference in bits means
Hash difference
Typical cause
Safe to delete?
0 bits + "Identical file"
Exact copy of the same file (backup, re-download)
Yes, keep any one
0-3 bits
Same photo saved at another size, quality or format
Usually; keep the highest resolution
4-8 bits
Light edits: brightness, filter, small crop, screenshot of the photo
Check side by side first
9-14 bits
Burst shots, the same scene seconds apart
Pick your favorite
15+ bits
Different photos with similar composition
Usually not duplicates
Tips for cleaning up a photo library
Back up first. Before deleting anything, make sure you have a copy of the library, especially when using a high sensitivity.
Keep the highest resolution. Messaging apps and social networks shrink photos, so the copy with the most megapixels is usually the original.
Use the CSV report to sort by folder or size in a spreadsheet when you clean up a large library.
Mirrored or rotated copies are not matched, because the fingerprint depends on orientation. Rotate them first if you need to compare them.
Frequently asked questions
How does the duplicate photo finder detect similar images?
It uses a perceptual difference hash (dHash). Each photo is shrunk to a tiny 9 × 9 grayscale grid and every cell is compared with its right-hand and lower neighbor, giving two 64-bit fingerprints that describe the photo's light and dark structure. Two photos are compared by counting how many bits differ (the Hamming distance); a small number means they look the same even if the files are different.
Can it find resized or edited copies?
Yes. Because the fingerprint describes the overall picture rather than the exact bytes, it still matches after resizing, saving at a different JPG quality, converting between formats, small brightness changes and slight crops. Heavy crops, rotations, mirroring and large edits change the fingerprint too much and will usually not match.
What sensitivity should I use?
0 bits finds only copies that look identical. 3 to 6 bits is best for the same photo in different sizes, formats or quality levels. 8 to 12 bits also groups near-duplicates like burst shots taken a moment apart, and above 14 bits you will start to see different photos with similar composition, so check those groups carefully.
Does this tool delete my photos?
No. Browsers do not allow a web page to delete files on your computer, and this tool never changes your files. It marks suggested duplicates so you can review them, then gives you the file list and a CSV report so you can delete them yourself in your file manager or photo library.
Are my photos uploaded?
No. The folder is read by your browser on your own device; photos are decoded and hashed locally with JavaScript. Nothing is sent to a server, so it works with private photos and even offline once the page is loaded.
Which file types are supported?
Any format your browser can decode: JPG, PNG, WebP, GIF, BMP, AVIF and ICO in all modern browsers. HEIC photos can only be read in Safari; in other browsers they are listed as skipped, so convert them to JPG first if needed.
How many photos can I scan at once?
Several thousand photos work well on a typical computer. Each photo is decoded once and only a small thumbnail and fingerprint are kept in memory, so the limit is mainly time: very large RAW-sized JPGs take longer to decode.
Are all the tools on FreeToolCabin tested for quality and accuracy?
Yes! Every tool on our site undergoes automated Q/A testing in real browser environments to ensure high precision, correct output, and zero errors. If you happen to spot a bug or have a suggestion, please contact us at [email protected] and our team will review it right away.