1) to display content: is it better to use a 3D animation representing
something like a folding notebook? Or using plane paginated tabs over
the background image?
Although the 3D animation might be really good eye candy, in my opinion at least, it gets real old real fast and is a sort of minimum 'lag' you're choosing to put in purposely. Meaning that until the animation completes the user won't see the information they're really trying to get at. This just gets annoying as time goes by. Just bring up the product info as fast as possible.
2) is it better using the real captured image as background, or
substitute it with a static product image. This way people won't be
forced to stay right in front of the label during the navigation.
Don't use the real captured image as the background unless you're going to start altering the color of your text based on the image background. Imagine the user's scans a primarily white boxed product. Now you have white text on white background making it hard to read. ( or a black box with black text ).
Instead, capture the image and allow the user to look at, switch to, the original picture if they want but when they read the info you provide it should be on a background that facilitates reading the info.
Also, as someone mentioned. Don't force the user to stand in front of the product. Allow them to snap a photo and fetch the info and look at it as they keep on walking. Much better experience than to have to stand in front of a something to read all about it.