Yesterday evening Semantic Web Atelier (SWAt) 11 took place in Graz, Austria. We had one great talk from Martin Höffernig about the architecture and implementation for a semantic validation service for multimedia metadata documents called VAMP. Martin had support from Werner Bailer and Georg Thallinger who also described the broader context of the work in detail.
So what does VAMP do? In short words: This validation service allows to verify the (partial) semantic correctness of an XML (especially MPEG-7) file. Take for example the temporal decomposition of a video file described in XML. You’ve scenes (longer sequences of video frames), which are grouping shots (shorter sequences of video frames). One semantic constraints is that no shot start point (time in the video) can occur before the start point of the parent scene. The same goes for end points: With the parent scene all child shots should end. However in distributed environment one can never be sure. Also automatic extraction routines also fail to detect the exact frame number, where shot boundaries are found. A semantic validation service provides means to detect such semantic errors and that’s what VAMP does.
More information can be found on the VAMP website and I think Martin Höffernig, Werner Bailer and Georg Thallinger are open to questions 😀
[Note this is a cross post from here]