Abstract | ||
---|---|---|
Current methods for the accurate recognition of instru- ments within music are based on discriminative data de- scriptors. These are features of the music fragment that capture the characteristics of the audio and suppress de- tails that are redundant for the problem at hand. The ex- traction of such features from an audio signal requires the user to set certain parameters. We propose a method for optimizing the parameters for a particular task on the basis of the Simulated Annealing algorithm and Support Vector Machine classification. We show that using an optimized set of audio features improves the recognition accuracy of drum sounds in music fragments. |
Year | Venue | Keywords |
---|---|---|
2005 | ISMIR 2013 | support vector machine,drum classification,simulated annealing,mel frequency cep- stral coefficients,simulated annealing algorithm |
Field | DocType | Citations |
Simulated annealing,Audio signal,Pattern recognition,Computer science,Drum,Adaptive simulated annealing,Speech recognition,Artificial intelligence,Discriminative model,Machine learning,Support vector machine classification | Conference | 2 |
PageRank | References | Authors |
0.46 | 9 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Sven Degroeve | 1 | 173 | 11.81 |
Koen Tanghe | 2 | 20 | 2.57 |
Bernard De Baets | 3 | 2994 | 300.39 |
Marc Leman | 4 | 34 | 4.71 |
Jean-pierre Martens | 5 | 91 | 11.81 |