The major challenge for any photobook creation is the following:
An user has for example 100 images from vacation. Some choose nearly the same object, if this is he case some will be better than others (sharpness, contrast, and so on).
For an automatic creation process of a photobook without the need for the user to "choose" images it is necessary, to identify images with similar objects. Some attributes can be Location (if GPS Data available in the EXIF data) or Time (if multiple images was taken within less than a minute), but this kind of information are not very offen available or accurate. To analyse images to identify similar images and find the "best of the image" with less risk of false positive is a challenge.
Comments
The major challenge for any photobook creation is the following:
An user has for example 100 images from vacation. Some choose nearly the same object, if this is he case some will be better than others (sharpness, contrast, and so on).
For an automatic creation process of a photobook without the need for the user to "choose" images it is necessary, to identify images with similar objects. Some attributes can be Location (if GPS Data available in the EXIF data) or Time (if multiple images was taken within less than a minute), but this kind of information are not very offen available or accurate. To analyse images to identify similar images and find the "best of the image" with less risk of false positive is a challenge.
There's a guy who's working in this area on HN.
I believe his web app is called Inapic.com
Thanks, seems to be the right think.