P/SCAIE has reached its limits. The Group dealing with "the Study of Content Analysis‐based Information Extraction in media production" urgently needs more broadcasters to join its work.
| What is it about? Automated information extraction refers to any process capable of extracting meaningful information (metadata) from audio-visual material using computer analysis processes. Examples of such processes are automated speech recognition, subject classification of spoken content, segmentation of newscasts into elementary news stories, face recognition and copy detection. These techniques are seen as attractive as they promise to reduce costs in production and archiving. In particular for: |
P/SCAIE reference model |
- IT‐based production;
- archive management applications;
- multi‐purpose and multi‐channel productions;
- news production.
First results
The Group currently consists of EBU Members from Italy, Belgium, Japan, Germany, Greece, Denmark and Russia. In the first year of the project these participants have managed to:
- define a contract for the use of EBU Members' audio-visual material for research & experimentation;
- define 13 requests for technology;
- organise the first International Workshop on Automated Media Exctraction, AIEMPro08 ;
- analyse an architecture for setting up a reference library;
- analyse a reference framework for tools evaluation.
How to join the work
If your organisation is interested to join the P/SCAIE work, then please contact its project manager: Jean-Pierre Evain .