Skip to main content

Multidimensional Information Access Using Multiple Modalities

Objective

In this project, we will develop new concepts and techniques to allow fast and natural access to multimedia databases. This will imply both the integration of available technologies in the domain of vocal interaction and multimedia information extraction, and the design and evaluation of novel technology for gesture designation and manipulation coupled with adequate graphical presentation logic.

Specific stress will be put in the project on producing:
(i) fully specified evaluation methods and protocols for multimodal interaction which can serve as a roadmap for the development of future devices and systems and
(ii) generic middleware components based on standardized exchange protocols.

The prototype which will be developed and evaluated thoroughly with various user groups, will serve as a test-bed for exploring new types of man-machine interactions and devices in the domain of information access applications.

Objectives:
The objective is to provide an integrated and comprehensive framework for the design of modular multidimensional/multimodal dialogue systems. It will be based on interaction scenario comprising constraints on the use of modalities, the relation with the underlying application and the graphical presentation procedures.
Human factor analysis will drive the development of the multimodal user interface.
Multilingual linguistic components will be provided and adapted by the partners. Besides, a haptic and tactile device for multidimensional interaction will be designed and integrated. The dialogue manager combines the information available from the application, the modality specific analysers and the graphical interface.
It will manage a user representation to integrate specific procedures for anaphora and deixis interpretation with robust dialogue management procedures.

Work description

The work within MIAMM will be achieved along the following lines:
* The MIAMM platform will be based upon the definition of a series of interaction scenarios for multidimensional interaction, comprising constraints on the use of modalities, the relation with the underlying application and the graphical presentation procedures that have to be implemented;
* A haptic and tactile device for multidimensional interaction will be designed and integrated within the platform after an initial evaluation phase of the current available technology.

This will be integrated into the prototype with an adequate visualization environment and gesture interpretation algorithms.
Fully multi-lingual understanding and generation components will be integrated within the MIAMM platform by incorporating and adapting the latest technologies in speech recognition and synthesis, and combining several alternative parsing technologies (statistical, automata, local syntax). Interface through the multimode interface language (MMIL) will result in language easily pluggable software modules.
A dialog manager will be designed to combine the information available from the underlying application, the language modules, the haptic device and the graphical interface.
Exchanging information by means of the MMIL language, it will manage its own user representation to integrate the latest scientific development concerning anaphora and deixis interpretation, as well as smooth dialog management to cope with both vocal and gesture interaction.
Human factor analysis will drive the development of the multi-modal user interface. For the haptic interaction human factor research has to find out for which purposes and actually how a person would use the haptic device. This will directly lead to semantic concepts related to haptic input and output gestures.

Milestones:
The projects milestones are after a requirements and specification phase of three months followed by the design and implementation phase planned for a duration of twelve months. After the second milestone a first prototype is available, that is evaluated and enhanced iteratively.
The final results are a working prototype and guidelines and evaluated best practice recommendations for the development of a multi-modal system.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE
Address
Domaine De Voluceau
78153 Le Chesnay
France

Participants (4)

CANON RESEARCH CENTRE EUROPE LIMITED
United Kingdom
Address
1 Occam Court. Occam Road, Survey Research Park
GU2 7YJ Guildford, Surrey
DEUTSCHES FORSCHUNGSZENTRUM FUER KUENSTLICHE INTELLIGENZ GMBH
Germany
Address
Erwin-schroedinger-strasse 57
67663 Kaiserslautern
NETHERLANDS ORGANISATION FOR APPLIED SCIENTIFIC RESEARCH TNO
Netherlands
Address
Schoemaker Straat 97
6060 Delft
SONY INTERNATIONAL (EUROPE) GMBH
Germany
Address
Kemperplatz 1
10785 Berlin