Face Tags in Photo Station

Started by Paul Barrett, February 25, 2017, 03:18:54 AM

Previous topic - Next topic

Paul Barrett

Aologies if this outside the usual scope of this form, and please move or delete if needed.

My Digital Asset Manager - Daminion, writes People tags to XMP:Person In Image.

My image publisher - Synology Photo Station - does not read that tag but relies on face tags applied by face tagging. For example:

Region Rectangle                 0.297949, 0.135652, 0.430639, 0.467826
Region Person Display Name       Florence Hackett

This is a bit of a longshot, but are there any Synology Photo Station users on here who have tried writing dummy data to these fields to hoodwink Photo Station into recognise the people without having to manually tag each one?

If so, how did you manage it?

Thanks

- Paul

StarGeek

You'll have to test it but try variations on this to see how Synology reacts

exiftool -struct -RegionInfoMP="{Regions=[{PersonDisplayName=Florence Hackett,Rectangle=0|,0|,0|,0}]}

Maybe remove the 0s (Rectangle=|,|,|,)or both 0s and pipes (Rectangle=,,,).  Or maybe some really small numbers, like changing the first 0s to 0.000001 and the second ones to 0.000002.  Really small numbers like that would just encompass part of a single pixel except on extremely huge images.

I would think that the first one would technically work, but I'm just guessing and you'll have to try it out. 

* Did you read FAQ #3 and use the command listed there?
* Please use the Code button for exiftool code/output.
 
* Please include your OS, Exiftool version, and type of file you're processing (MP4, JPG, etc).

Paul Barrett

Thanks StarGeek

I had already tried zeroes with no luck.  I could try some small values.  But, now I have thought about it some more i think this is too complex to tackle.

Reason: The person's name that needs to go into the command would require me to run the command 200 times - error prone. The name could be retrieved from the keywords field where there will be entries that take the form "People|"<name>", "

But many of the images have multiple people in multiple keywords, mixed up with keywords for Events, Places, Holidays, Animal, Miscellaneous.  Parsing those fields to isolate the people would be tough and we can't use the same coordinates for multiple people on one photo so I'd need to increment the coordinates.

I think that's a mountain that I will not tackle after all:)

Paul