https://github.com/cgoldberg/dupes
detect visually similar images
https://github.com/cgoldberg/dupes
Last synced: 4 days ago
JSON representation
detect visually similar images
- Host: GitHub
- URL: https://github.com/cgoldberg/dupes
- Owner: cgoldberg
- License: mit
- Created: 2019-05-18T16:22:14.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-05-18T16:24:24.000Z (over 6 years ago)
- Last Synced: 2025-01-11T15:28:42.209Z (9 months ago)
- Language: Python
- Size: 2.93 KB
- Stars: 0
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# dupes
Corey Goldberg (2019)
License: MIT
* https://github.com/cgoldberg/dupes
----
## Detect visually similar images.
`dupes.py` recursively scans a directory tree to find visually similar images. It identifies similar images by comparing a wavelet hash of each. A fingerprint is calculated using a perceptual hashing algorithm (wavelet hash) from the `ImageHash` library. Images that hash to the same value are considered visually similar. Paths to the suspected duplicate images are printed to the console.
#### Wavelet Hashing:
* https://fullstackml.com/wavelet-image-hash-in-python-3504fdd282b5
#### Requirements:
* Python3
* ImageHash
* `$ python3 -m pip install imagehash`
* https://pypi.org/project/ImageHash/
* Note: `ImageHash` package installation depends on: `numpy`, `Pillow`, `PyWavelets`, `scipy`, `six`
#### Usage:
* `$ python3 dupes.py [dir]`
* Note: if no directory is specified, scan starts from the current directory