Use Case: Recommendation across different media types

use case related to issue 6064

Introduction

The multimedia resources (video, audio, image, etc.), increase rapidly in the Internet. How should these resources be used to provide useful services and attract more and more interests?

Currently, multimedia content providers (youtube, iptv content provider, etc.) have their specific or proprietary metadata formats (see [http://www.w3.org/2005/Incubator/mmsem/XGR-vocabularies/ mmsem multimedia vocabularies). This is one of the key practical problems faced by service providers.

Harvesting and parsing metadata should be greatly facilitated if a unified annotation format were used to publish metadata or map metadata to from different sources.

Motivating Example

From the perspective of a service provider, the following example can be seen as typical scenario. What should be done to offer more opportunities?

IPTV will be the next generation of Television and will gradually replace traditional broadcast. In IPTV, the television set will be connected to the Internet and provide access to traditional TV programmes, but also other multimedia resources (youtube, etc.).

In this new paradigm the challenge will be for service providers to be discovered to offer their services, and for users to find content of their interest. One can foresee different possible approaches to search and retrieve /access content:

• Service providers can recommend access to their services by publishing electronic programmes guides either pushed to the user (e.g. via a dedicated set-top-box) or accessible via a portal. Such guides will exploit metadata in different formats adapted to the type of query. Another option is to publish raw metadata for use by third parties (acting purely as recommendation service providers) such as portal developers, manufacturers or search engines. In this case, the use of a common annotation format would be of great benefit.

• However, some content providers on the web do not support metadata access API. In this case, a recommendation service provider has to do web page extraction to infer the possible metadata models, which leads to very poor reusability and scalability. Such service providers should be encouraged to use a simple unified annotation format to facilitate data mining to their own benefit for a larger access to their services.

Based on the recommendations and solutions provided by this group, the service providers will be able to design and implement the most appropriate metadata and algorithms for their respective services.

Related Resources

• Multimedia Vocabularies on the Semantic Web, W3C Incubator Group Report 24 July 2007. MMSEM XG report on current multimedia vocabularies.
• Video Video Recommendation Based on User Preference on the Web, by Soohong Daniel Park, Cheolju Hwang and Glenn A. Adams, Samsung Digital Media R&D Center. One of the possible recommendation service examples.