A Music Retrieval Method Based on Distribution of Feature Segments

In this paper, we propose a music retrieval method based on the distributions of features in the music. In common music retrieval methods, if several features are similar between the query and the retrieval target, the retrieval systems return that the query is similar to the retrieval target. However, a problem is that several features in the music are ignored. If the other features in the query and the retrieval target are quite different, the query and the retrieval target should be treated as different types of music. Therefore, we calculate the importance of each feature in the music. Then, we compare the importance of features between the query and the retrieval target, and we can retrieve the music without ignoring the importance of several features. In our experimental evaluation, we can confirm that our proposed system has better accuracy than the baseline method.