Limitations of the MPEG-7 Generic DS: Reorganizing the Syntactic/Semantic DS’s

In this document, we propose some modifications to the MPEG-7 Description Scheme (DS) [1] in order to enrich the structure of the Syntactic DS and Semantic DS by addressing some functionalities for semantically characterizing segments and for highlighting and ordering key-items in a multimedia document. In our opinion, the Generic DS and in particular the syntactic DS can demonstrate some weakness in describing hierarchically organized documents. In other words, even if it is enunciated [1] that the Syntactic DS should act as the Table of Contents for the multimedia document being described, the description of the document temporal structure seems complicated. Therefore we start our discussion by implementing the ToC DS part of our old MPEG-7 proposal (the ToCAI DS [3]) using the MPEG-7 Generic DS [1]. In our opinion, due to the simpler structure of the ToC DS, this implementation allows to show the complexity of the MPEG-7 DS. For overcoming such a problem, we propose a simble extension of the Syntactic DS of the MPEG-7 Generic DS in order to handle semantic aspects of each segment directly at the Segment DS level. Another issue that we analyze in this document is a possible extension of the MPEG-7 Generic DS for the inclusion of some important functionalities: the capability to (1) highlight description items (e.g. images, sounds, events, objects etc.) most relevant to the purpose for which a certain content description of a multimedia (MM) document has been created and (2) the capability of description information ordering . In other words, due to a possible large amount of description items, an entity who will create descriptions of multimedia (MM) documents, according to MPEG-7 specification (i.e. a description provider), shall highlight certain items most representative for the kind of document being described in order to facilitate user queries. Besides we consider the need of providing users with ordering mechanisms a very relevant issue for MPEG-7. Such ordering mechanisms are derivable from descriptors (e.g. a set of key – frames ordered on the basis a color descriptor or a set of sounds ordered by means of a loudness D). However a possible large variety in the types of descriptors composing a description could lead to a consequent high number of ordering criteria to arrange description items. Therefore we propose that the description provider should also select which set of descriptors should be combined to order a subset of description elements (e.g. key frames, events etc.) most pertinent to the MM document being described. The document is organized as follows: in Section 2 we explain the motivations for a representation of the ToC DS based on the MPEG-7 Generic DS. In Section 3 after a brief overview of the ToCAI DS, we present the implementations of the ToC DS according to the MPEG-7 Generic DS specifications; we also suggest in this sections some changes to the current specifications to better handle the ToC DS functionalities. Section 4 provides an example of implementation of such a DS. In Section 5, we explain, after a quick overview of the current Generic DS, the motivations behind the proposal for adding highlighting and ordering functionalities. In Section 6, we show the structure of the DS that enable these functionalities. In Section 7, we give an example in order to clarify the concepts of key-items and ordering keys. Finally in Section 8, we provide a brief summary of the contribution.