Skip to main content

Python 3 CLI application for finding visually similar images

Project description

usage: imagedupes [-h] [-a ALGORITHM] [-d DIRECTORY] [-D DATABASE] [-l]

[-o OPTIONS] [-p PROGRAM] [-r] [-R] [-s HASH_SIZE]

Finds visually similar images and opens them in an image viewer, one group of matches at a time. If no options are specified, it defaults to searching the current working directory non-recursively using a perceptual image hash algorithm with a hash size of 8, opens images in the system default image handler (all members of a group of matches at once), and does not follow symbolic links or use a persistent database.

optional arguments:
-h, --help

show this help message and exit


Specify a hash algorithm to use. Acceptable inputs: ‘dhash’ (horizontal difference hash), ‘dhash_vertical’, ‘ahash’ (average hash), ‘phash’ (perceptual hash), ‘phash_simple’, ‘whash_haar’ (Haar wavelet hash), ‘whash_db4’ (Daubechles wavelet hash). Defaults to ‘phash’ if not specified.


Directory to search for images. Defaults to the current working directory if not specified.


Use a database to cache hash results and speed up hash comparisons. Argument should be the path to which you want to save or read from the database. Warning: runnning the program multiple times with the same database but a different hash algorithm (or different hash size) will lead to missed matches. Defaults to no database if not specified.

-l, --links

Follow symbolic links. Defaults to off if not specified.

-o OPTIONS, --options OPTIONS

Option parameters to pass to the program opened by the –program flag. Defaults to no options if not specified.

-p PROGRAM, --program PROGRAM

Program to open the matched images with. Defaults to your system’s default image handler if not specified.

-r, --recursive

Search through directories recursively. Defaults to off if not specified.

-R, --raws

Process and hash raw image files. Note: Very slow. You might want to leave it running overnight for large image sets. Using the –database option in tandem is highly recommended. Defaults to off if not specified.

-s HASH_SIZE, --hash_size HASH_SIZE

Resolution of the hash; higher is more sensitive to differences. Some hash algorithms require that it be a power of 2 (2, 4, 8, 16…) so using a power of two is recommended. Defaults to 8 if not specified. Values lower than 8 may not work with some hash algorithms.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

imagedupes-1.2.6.tar.gz (19.9 kB view hashes)

Uploaded Source

Built Distribution

imagedupes-1.2.6-py3-none-any.whl (8.4 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page