idealo / imagededup

😎 Finding duplicate images made easy!
https://idealo.github.io/imagededup/
Apache License 2.0
5.16k stars 456 forks source link

Not working on the below type of images #80

Closed vyaslkv closed 4 years ago

vyaslkv commented 4 years ago

Please help how could I find duplicates in such type of images (have tried both the methods cnn & hash)

test2_new

datitran commented 4 years ago

Uhmm can you provide us with a working example like some samples of your images and also the code that you use? Otherwise we are not able to help...

vyaslkv commented 4 years ago

Thanks @datitran for responding

these are the images & I am using the standard code which is given in your examples in this repo

test1 test2 test3 outside
vyaslkv commented 4 years ago

Like in this I want to detect last two as duplicates or identical

tanujjain commented 4 years ago

@vyaslkv If I understand you correctly, you ran the standard phash readme code and got a dictionary back where no duplicates were found?

You can try to vary the corresponding threshold parameters for hashing and cnn methods to see which threshold works best for you. You can additionally use the score attribute to see the extent of difference/similarity between the pair of images you ascertain to be duplicates.

Please refer the documentation for these: https://idealo.github.io/imagededup/

vyaslkv commented 4 years ago

Thanks, @tanujjain Let me check again. (Really Thanks to all of you for the quick help I really appreciate)

tanujjain commented 4 years ago

@vyaslkv Gave your images a try with default settings for phash and cnn. The images were found to be duplicates in both cases as expected. So, closing the issue.