Hello,
I am investigating the potential use of Vuforia for a client. For this I implemented the demo version of Vuforia into a Unity project. There, I have a variety of image targets on one hand and an Unity object on the other. I want to track the images and place the object at their position, no matter which image it is (After that I actually change some UI elements in the object according to the image that was tracked). All of this I want to do by using a device database.
With this I have two issues:
1. In your example for "device database driven image tracking", every possible image needs to have one individual and active game object in the scene. So, if I want to track 250 different images, this would mean that I need 250 active game objects in the scene all the time, all running the "image target behaviour". Did you consider the performance implications of such a setup? What if I want to track 1000 different images? Do I need to create 1000 individual and active game objects? I was wondering especially because in the cloud based example, the tracking behaviour was implemented by using a customizable C# script, whereas for "device database driven tracking", all important scripts are hidden behind the .dll protection.
2. Am I missing something or is there no easy way to have a script that registers itself in Vuforia to retrieve notification if a new object was tracked or if tracking was lost, including all the parameters such as: object scale, object rotation, tracked object name etc. It seems to me that Vuforia was designed in a way for extremely easy use by reducing developers ability for more customized solutions. Hopefully I am wrong in this regard!
In an ideal world I would like to just say:
IF there was a new object detected, THEN instantiate the game object that should be placed on all image targets and place it on the targets position. After that change some values on the target based on the object that was tracked. IF the targets tracking position was lost completely, THEN destroy the instantiated game object. Is there a way to make this possible?
Thanks a lot for you support,
Danish