The way Vuforia works is really really simple, but what matters here is do you wanna dynamically translate any book or you wanna stay with the legacy way to AR.
And by that I mean, do you wanna be able to translate any book by reading its content using OCR algorithms, search internet for relevant 3D content and if available, show that on paper as a 3D representation. OR your targets id rather simple and limited to a bunch of books and a bunch of target images, if that is the case things are gonna be really really simple.
The first approach is gonna be a pain, but for the second one you ll just need a bunch of target images and a 3D representation/model for each image, you can than map them using vuforia engine. If you wanna go for a larger database, I would suggest to categories them and keep most of the data in the cloud
Hope this helped…😃