Project: Extraction of semantic information from audio signals
- project duration
- 01-OCT-05 – 30-SEP-10
- The proposed research is about the development of new algorithms contributing to the more reliable extraction of semantic information from audiostreams in which several sound sources are simulaneously active. The algorithms will be applied to the following domains: the extraction of melodic and rhythmic information from polyphonic music and the recognition of speech in the presence of interfering background sounds.