Summary
The objective of the project is to provide technical devices and systems for digital preservation of all types of audio-visual collections. The aim is to build-up preservation factories providing affordable services to all kinds of collections owners to manage and distribute their assets. The way to achieve this goal is an integrated approach, to produce sustainable assets with easy access for larger exploitation and distribution.
At present, there are many unsolved problems of digitisation, automation, metadata extraction, storage, network bandwidth, secure interaction, access and end-user delivery. Partial solutions exist, but in general they are not robust, scaleable or affordable - and definitely not integrated end-to-end within a sustainable commercial model. Prestospace is funded as an Integrated Project under European Commission 6th FP with a budget around 3.3M euros. It starts from February 2004 and runs for 40 months. Prestospace is co-ordinated by Daniel Teruggi, Institut national de l’audiovisuel (Ina), France.
Contact: Hamish Cunningham (PI).
Project Objectives:
The overall scientific and technological objective is to develop an integrated approach to audiovisual preservation and access, to produce sustainable assets with high cultural value and potential commercial use.
This objective will be achieved through new technology and processes, and will be implemented as facilities that will provide all holders of audiovisual material with an integrated affordable preservation service.
PrestoSpace will provide:
Our Prestospace area of work is metadata and access. Content analysis providing high-level information and structure description of the documents particularly contributes to the easy access and exploitation of the audiovisual assets. Thus the recovery of the initial legacy information should be combined with selected analysis algorithms for producing time-based content description.
Part of the analysis can be made in real-time during the transfer to digital. Other extraction processing can be devoted to retrieval and access purposes and can be done as complementary service off-line. Authoring tools will be provided according to users requirements to enhance access and reuse of the preserved collections.
Speech recognition and transcription need to be extended to new languages according to simplified learning processes. Then the results provide important knowledge for building a semi-automatic indexing system.
Most of these research and developments will enable new classes of industrial services. The economic results should benefit from research with a European and international basis, rather than the proprietary development and use of the technologies. To stimulate the largest dissemination of results, when possible, the outcomes will be delivered into the open source environment.
Our Role
Research in Human Language Technologies has demonstrated the effectiveness of Information Extraction (IE) techniques. Some of them have already been integrated in applications for digital libraries, multimedia indexing, automatic news classification, Web hyperlinking and medical informatics. In Prestospace we are working on innovative Information Extraction (IE) methods for metadata and access to multimedia content. This is a continuation of our work in an earlier project: MUMIS.
The Information Extraction technology will be integrated in:
More specifically we are working on:
Selected Publications
Funding: