2.7 Encoding and Decoding of MPEG-7 Documents for Delivery-Binary Format for MPEG-7

---

2.7 Encoding and Decoding of MPEG-7 Documents for Delivery—Binary Format for MPEG-7

MPEG-7 descriptions may in many cases be generated automatically. We have seen low-level descriptors that contain histogram-based values of size 256; a complete video may be automatically decomposed into its segments, and for each segment different low-level descriptors may be available. The resulting XML document may thus be verbose and is not suitable for consumption in a constrained and streamed environment. To overcome the lack of efficiency in this textual XML, MPEG-7 Systems defines a generic framework to facilitate the delivery and processing of MPEG-7 descriptions: the BiM (binary format for MPEG-7). It enables the streaming and compression of any XML document.

It is important to note that BiM coders and decoders can deal with any XML language. Technically, the schema definition of the XML document is processed and used to generate a binary format. This binary format has two main properties. First, because of the schema knowledge, structural redundancy (element name, attribute name, and so forth) is removed from the document. Therefore, the document structure is highly compressed (up to 98%). ^[13] Second, element and attribute values are encoded according to some dedicated codecs. A library of basic datatype codecs is provided by the specification (e.g., Zlib). Other codecs can be plugged easily using the type-codec mapping mechanism defined in the standard.

One of the main technical advantages of the BiM binary encoding process is that it can be guided by schema information. In BiM, the schema is known both by the encoder and the decoder. The binary format is deduced from the schema definition. Thus, there is no need to define coding tables or a specific encoding mechanism.

2.7.1 Required Software

The current version of the BiM Reference Software may be obtained from http://www.expway.fr/mpeg/bim/, though it is password protected by MPEG). An older version is included in the Reference Software available at http://www.lis.e-technik.tu-muenchen.de/research/bv/topics/mmdb/e_mpeg7.html (see als Section 2.9). Encoding details may be found in the document ISO-IEC 15938-1.

Each document can be transmitted in one or more pieces (called access units, the minimal unit that has a time stamp for decoding). At the lowest level of granularity, each attribute value or document leaf can be modified to allow a minimal transmission in case of a minimal change in the sent document.

Example: Consider the following MPEG-7 document describing a 1.5-minute-long video with two VideoSegments: NarrationVS (the first 15 seconds) and CaptureVS.

 <Mpeg7>   <Description xsi:type="ContentEntityType">     <MultimediaContent xsi:type="VideoType">       <Video >         <MediaTime>           <MediaTimePoint>T00:00:00</MediaTimePoint>           <MediaDuration>PT1M30S</MediaDuration>         </MediaTime>         <TemporalDecomposition gap="false"             overlap="false">           <VideoSegment >             <MediaTime>               <MediaTimePoint>T00:00:00<                   /MediaTimePoint>               <MediaDuration>PT0M15S</MediaDuration>             </MediaTime>           </VideoSegment>           <VideoSegment >             <MediaTime>               <MediaTimePoint>T00:00:15<                   /MediaTimePoint>               <MediaDuration>PT1M15S</MediaDuration>             </MediaTime>           </VideoSegment>         </TemporalDecomposition>       </Video>     </MultimediaContent>   </Description> </Mpeg7> 

A possible partition of the example document is shown in Exhibit 2.16. The global information of the complete video (RootV) is put in the first access unit, and then we encode the information of the temporal decomposition and the video segments NarrationVS and CaptureVS in the second access unit. Using two access units leads to the coding scheme as depicted in Exhibit 2.16. The document is here displayed as a tree, where nodes represent the elements. For a better readability, the names of some of the elements are not shown.

Exhibit 2.16: Encoding an Extensible Markup Language (XML) document in pieces.

The advantage of encoding a document in several pieces is that the pieces can be delivered separately to the client. It is not required for the decoder to download (and keep in memory) the entire XML file before being able to process it. This ability for separate delivery can reduce both the memory required at the terminal side and the consumed bandwidth.

^[13]Niedermeier, U., Heuer, J., Hutter, A., Stechele, W., and Kaup, A., An MPEG-7 tool for compression and streaming of XML data, in Proceedings of the IEEE International Conference on Multimedia and Expo, Lausanne, Switzerland, August 26–29, IEEE CS Press, 2002, pp. 521–524.

---