I started the implementation of the high-level interface for the CASIA-SURF database. It's the data that is used for the ChaLearn PAD competition I'm supposed to work on.
So I thought this might be of interest to you to discuss that here (you may have a look at the low-level package
In particular, this database is image-based (with mutiple modalities) and I was wondering if an ImagePadFile Class would make sense, and how it would impact the whole framework (if it does). It also made me think about this whole FrameContainer stuff - which I still don't get, sorry ;) ... For the moment, I will stick to what has been done and consider an image as a video of one frame.
Anyway, any input/comment is welcome !