new media for a new millennium
The objectives can be characterised as:
These are illustrated in the figure below.
In this context the AiiA Lab will help in the construction of the so called NAL, by which semantic entities will be extracted with the use of an inference machine. Our role in this task is to identify the target facts that can exists within the content based video annotation and give specific notions that will be used.
Media Annotation is a very time consuming process if we want to do it manually. Our lab will propose algorithms with which semi-automatic annotation can be realized. Among them the most popular are face detection, face tracking and shot detection. Other as well will be used and also developed within this project. The results of those algorithms will then be used to annotate multimedia essence files in an XML MPEG-7 compatible format.
Narrative understanding is a high semantic processing task. It is the one which will identify the narrative structure of a multimedia subject. In this context the AiiA lab will provide knowledge of media annotation which will construct in later time the bases for extracting narrative understanding annotion.
The Content understanding task will identify what is happening within a multimedia clip. AiiA lab will provide algorithms to do so. Those algorithms will combine low level features in order to provide more significant (human perception based) features.