4

From a group photo, I would like to create an image that identify each person with their names. My first idea was to create a silhouette image from the original one, put a number over each face, and present a list of number-name. I could do that using GIMP, for example.

Is there any more intelligent way of doing so?

Windows Photo Gallery, for example, has a tool for tagging people. Is it possible to produce a new image with tag information from the original one?

  • Don't know the answer, but I hope someone does. It would be a useful tool. I'd bet good money that most intelligence agencies have this. The first step would be to identify all the people and a number of apps have tools for that. KPhotoAlbum on Linux recently added automatic face recognition and tagging. The problem is that systems like this probably don't save the coordinates where each match occurred, so there would be no direct way of knowing where to graphically tag/annotate the image. – Joe Jul 28 '15 at 22:30
  • @Joe, face recognition would be nice but I would be happy to have a tool to create group photo silhouettes tagged with people's names even if I had to enter the names manually. – toliveira Aug 5 '15 at 10:43
  • I asked on the KPA list. Here are a couple of comments which indicate how you might proceed. Basically, KPA already has the infrastructure to mark people (or whatever you want). This has even be added before the face management, by the positionable tags. You can mark areas on images and add a tag to them. So ... we already do have this functionality! Cont... – Joe Aug 6 '15 at 17:40
  • Second comment: Just as a quick idea this could probably be implemented as an independent python script. The KPhotoalbum python library (although being quite slow) can parse all relevant data from the database file, which you could then use in a python program to generate the segment-images. – Joe Aug 6 '15 at 17:43
1

If you're prepared to do some of your own development then you could maybe start with http://www.face-rec.org/.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.