M-MUSICS

摘要: Accurate voice humming transcription and efficient indexing schemes are essential for a large-scale humming-based music retrieval system. Although many researches have been done to develop such schemes, their performances not still satisfactory. In our previous works, we proposed (i) new query scheme [4], (ii) popularity-adaptive structure called FAI [6] fast retrieval, (iii) semi-supervised relevance feedback reformulation based on genetic algorithm [7] in order improve efficiency. this demonstration, extend efforts mobile environment prototype system M-MUSICS. Our focus implementation includes versatile user interface easy querying browsing typical device as PDA phone satisfactory performance wireless environment. We report some of the results.