Selection of extraction algorithms (A/S, UPF)

Best suited algorithms to targeted applications features will be selected:

In the case of the Web monitoring system, it will require the algorithm to process a large amount of data in a very little time.
The retrieval features of the Sound Palette will require more precise and flexible processing

Research on feature extraction will also aim at developing techniques for automatic elementary sound extraction from compound mixtures, performing analysis tasks such as time segmentation, source separation, multiphonic pitch detection etc.

Although this part is more speculative, related output will be applied as much as possible to high-level feature extraction from multiphonic music pieces, in coordination with WP2.2.

Page remise à jour le 14 Janvier 2001