Abstract
In the research of images or videos, there are some complete systems to support query functions in different features. However, in music there is no system to do such work. In this paper, we propose four indexes to meet the above need. Based on the index of suffix tree, we dev elop the first two indexes, Multi-Features Suffix Tree and Independent Suffix Trees. We discover different problems of search functions in each specific index. Then we develop the another index, Twin Suffix Trees, to improve these problems. However, they lack of scalability when amount of music data increases. Therefore we propose the forth index, Grid-Twin Suffix Trees, to provide scalability and flexibility for large amount of music data. In each index, we could request different search functions, like exact search and approximate search, on different music features, like melody, rhythm or both. In experiments, we would compare the performance of different search functions in each index. Therefore, system manager could choose any one of these indexes according to the need of system and the amount of music data