BROWSE
Detail
Conference Paper
A Music Similarity Function Based on the Fisher Kernels
Cited - time in 
Share 
Share
- Authors
- Jin S. Seo, Nocheol Park, Seungjae Lee
- Issue Date
- 2012-06
- Citation
- International Symposium on Computer Music Modeling and Retrieval (CMMR) 2012, pp.429-436
- Language
- English
- Type
- Conference Paper
- Abstract
- Music-similarity computation is an essential building block for browsing, retrieval, and indexing of digital music archives. This paper presents a music similarity function based on the Fisher-vector representation of the spectral features extracted from a song. The distance between the Fisher vectors of two songs is used as the similarity of the two songs. The Fisher vector has a closed-form representation and can be readily incorporated with simple vector distance measures. Experimental results show that the Fisher-vector representation of the auditory features is promising for the music-similarity computation.
- KSP Keywords
- Building block, Closed-form representation, Music Similarity, Similarity computation, Similarity function, auditory features, digital music, distance measure, fisher vector, spectral features, vector representation