I am working in an Augmented Reality project based on ARToolkit, and other libraries like openCV. I am using Collada geometry to represent the virtual information on the scene. The image below shows a little about it.
However, a big problem occurs when the 3D objects are complex enough, and the reading of those become very hard.
Then, I would like to know how can I read complex xml objects (.dae) faster. Is there any binary format of Collada geometry?
Thanks for now, and sorry for my English mistakes.