Main Menu

Home
Coordination
Research & Integration
Dissemination
Community
News
Events
Links

Workpackages

WP2: Evaluation, Integration and Standards
WP3: Visual Content Indexing
WP4: Content Description for Audio, Speech and Text
WP5: Multimodal Processing and Interaction
WP6: Machine Learning and Computation Applied to Multimedia
WP7: Dissemination towards Industry


Logo IST
E-teams

The purpose of E-teams is to stimulate scientific collaboration between different MUSCLE partners. As the success of such collaborations is best measured in terms of scientific output, E-team activity will be evaluated first and foremost in terms of the number of joint research papers submitted to international journals and conferences. Other indicators of effective collaboration, such as joint patents or jointly developed software, will also be taken into account, albeit to a lesser extent.

Size and Scope

There is no upper bound on the number of partners involved, but as collaboration typically works best in smallish groups, it is expected that most E-teams will comprise between 4 and 8 partners. Also, research interests of individual researchers might change and and E-team membership will evolve accordingly. Finally, non-MUSCLE researchers can join an E-team but they will not be eligible for funding. An E-team's research goals and topics should be in line with MUSCLE's scientific objectives as defined in the Joint Programme of Activities (JPA).

Evaluation criteria

The value of E-teams in gauged in terms of their scientific contributions to MUSCLE's objectives (as defined in the Joint Programme of Activities). For that reason evaluation will be based on the following criteria:

  • Number of joint papers in leading international journals and conference proceedings

  • Indicators of successful collaboration (such as joint patents, jointly developed software, etc) and contributions to MUSCLE's objectives as defined in Joint Programme of Activities (in particular WPs and Challenges).

  • Evidence of collaboration and activity on publicly accessible documents, such as the MUSCLE and E-team webpage, MUSCLE reports, etc

Funding

As the E in E-teams suggests, collaboration will primarily be conducted through electronic means. However, the Network will support active E-teams through grants for mobility and collaboration. Eligibility for such grants will depend on the "maturity" of the E-team. In its start-up phase, an E-team is expected to clearly define its research goals and start collaboration mainly through electronic means. Moreover, such teams are encouraged to organize focus meetings in the wake of international conferences or workshops (e.g. as most members of an E-team will share similar interests, they are very likely to attend the same international conferences). E-teams can then request funding to pay for the additional costs thus incurred (e.g. renting meeting room, hotel accommodation for an extra night, etc).

As soon as an E-team can demonstrate activity in terms of the afore-mentioned evaluation criteria, it will be eligible for more substantial funding:

  • short exchange visits (e.g. 3-5 days) for senior or junior staff;

  • longer exchange visits (e.g. 1 month) for PhD students or postdocs;

  • invitation of external experts, etc.

It needs to be stressed that approval of funding is not automatic, the SC will base its decision on both the merits of each individual request and current priorities of the Network.

Choosing Features for CBIR and Automated Image Annotation
Statistical analysis of visual processes
Multimodal Processing and Multimedia Understanding
Mutlimodal interfaces on mobile devices
Audio-Visual Automatic Speech Recognition (AV-ASR)
Visual Saliency
Kernel methods in dynamic classification problems
Active Semi-Supervised Learning
Dynamic Texture Analysis and Detection in Video
Semantic from Audio and Genre Classification for Music
Integration of structural and semantic models for multimedia metadata management
Person detection, recognition and tracking
Shape modelling
Unsupervised segmentation and classification of multichannel data
Semantic from Audio: features, perception and synthesis