Pixel based duplicate image detection

Alison Kim 6 months ago in Digital Asset Management updated 5 months ago 4

As an administrator it would useful to perform an annual scan of all of our images to detect duplicates based on the pixel content. This would allow us to do yearly audits and find duplicate images that already exist in the DAM, regardless of filename. 

File Types Versioning Maintenance

About the compression: when you apply a compression on an asset, or color management, the asset is changed. The pixel values in the picture are different, so those are not intended to be identified as duplicates. If you would like that functionality, can you add an new voice ticket about this use case on its own, as this is a completely different ask/topic.

The ability to have more smartness related to duplicate detection we do see as valuable. Therefore i'll qualify this request. You will also find other qualified voice tickets about the topic which may be relevant.

We would like to clean up our system. We are in the process of migrating multiple libraries of photos. These photos were not subject to the checksum test because we didn't upload them manually. 

For manual uploads we've also found that the checksum doesn't work if the photos have different filenames. As you can see in the screenshots below, I was able to upload two identical photos that have the same checksum "1140667594." The third screenshot is the original photo that I downloaded from. We would also like to prohibit the uploading of compressed versions of photos if the original already exists. 

In Review

The DAM is able to identify files based on the exact same content using the checksum it calculates. This will identify true duplicates. Were you referring to true duplicates or to similar content? Can you elaborate what you are trying to accomplish with the yearly audit, is this a cleanup of the system, a report on the % of duplicate content, or something else?