[Digikam-devel] Re: Needing ideas for Rosedu's Open Source Development Course

[Michael G. Hansen]

On 01/05/2011 01:02 PM, Michael G. Hansen wrote:
> On 01/05/2011 12:36 PM, Gilles Caulier wrote:
>> 2011/1/5 Gabriel Voicu<ping.gabi at gmail.com>
>>
>>> Hi Gilles,
>>>
>>> Thank you very much for this long answer! I will speak today with the
>>> organisers and see what they have to say.
>>>
>>> Also, be sure that I was not thinking in starting a junior job project
>>> inside digikam's core, because it's under heavy development for the 2.0
>>> version. The ideal project would be modifying a GUI for a plugin(album or
>>> KIPI, doesn't matter).
>>> Do you have in mind a plugin that should need GUI refactoring? If yes,
>>> please let me know, because I couldn't find any example on Bugzilla.
>>>

We also still don't have any code that saves people tags with regions to 
files. We discussed some possible options about using RDF etc. So here 
one could write code that takes a face tag with region coordinates, and 
writes it into XMP via libkexiv2, and can also produce a face tag from 
the XMP. Here, one would need to know the representation of face tags 
inside digikam, and then spend some time getting save/load code to work 
using test based development, without touching any digikam code. Once 
the code is done, either the students or a digikam developer would 
integrate it into the image metadata load/save routines of digikam.

Document containing research on face tag representations:

http://websvn.kde.org/branches/extragear/graphics/digikam/core/project/documents/ImageAnnotation.odt?view=log

I think at the last sprint, we were most favorable of the scheme used my 
Masahide Kanzaki:

http://www.kanzaki.com/docs/sw/img-annotator.html

Required knowledge would be RDF/XMP, and how libkexiv2 represents it.

Michael