[
https://issues.apache.org/jira/browse/IMAGING-112?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13962951#comment-13962951
]
Xavier Dury commented on IMAGING-112:
-------------------------------------
Another method (to remove all the metadata) or even a flag is fine for me.
Currently, I export each image to a raw format which has no metadata associated
and then I compute hashes... but the drop in performance is dramatic. Another
way would be to provide a way to isolate/extract the image raw data bytes and
compute the hash directly from that... but I don't know if that is possible.
> JpegIptcRewriter.removeIPTC() does not remove all metadata
> ----------------------------------------------------------
>
> Key: IMAGING-112
> URL: https://issues.apache.org/jira/browse/IMAGING-112
> Project: Commons Imaging
> Issue Type: Improvement
> Affects Versions: 1.0
> Reporter: Xavier Dury
> Attachments: ImageTest.java
>
>
> I'm trying to compute hashes from my images to detect duplicates.
> To do so, I strip every metadata (like keywords I may have added) then
> compute a CRC32.
> My problem is that JpegIptcRewriter.removeIPTC() does not remove all metadata
> as there is always a difference of 18 bytes (containing "Photoshop 3.0").
> So could it be possible to remove those 18 bytes as well in removeIPTC() so
> that I can compute hashes?
> Thanks,
> Xavier
--
This message was sent by Atlassian JIRA
(v6.2#6252)
