M. Leimeister, “Feature Learning for Classifying Drum Components from Nonnegative Matrix Factorization,” in Proc. AES Convention 138, May 2015, pp. 9261:1-9. [Online]. Available: https://aes.org/publications/elibrary-page/?id=17685
Leimeister M. Feature Learning for Classifying Drum Components from Nonnegative Matrix Factorization. In: AES Convention 138. Audio Engineering Society; 2015.pp. 9261:1-9. Available from: https://aes.org/publications/elibrary-page/?id=17685
@inproceedings{Leimeister2015_17685,
author = {Leimeister, Matthias},
title = {{Feature Learning for Classifying Drum Components from Nonnegative Matrix Factorization}},
booktitle = {AES Convention 138},
pages = {9261:1-9},
url = {https://aes.org/publications/elibrary-page/?id=17685},
year = {2015},
month = may,
publisher = {Audio Engineering Society},
address = {Warsaw Poland}
}
TY - CPAPER
TI - Feature Learning for Classifying Drum Components from Nonnegative Matrix Factorization
AU - Leimeister, Matthias
T2 - AES Convention 138
SP - 1
EP - 9
M1 - Paper 9261
PY - 2015
DA - 2015/05/06
UR - https://aes.org/publications/elibrary-page/?id=17685
PB - Audio Engineering Society
CY - Warsaw Poland
LA - en
AB - This paper explores automatic feature learning methods to classify percussive components in nonnegative matrix factorization (NMF). To circumvent the necessity of designing appropriate spectral and temporal features for component clustering, as usually used in NMF-based transcription systems, multilayer perceptrons and deep belief networks are trained directly on the factorization of a large number of isolated samples of kick and snare drums. The learned features are then used to assign components resulting from the analysis of polyphonic music to the different drum classes and retrieve the temporal activation curves. The evaluation on a set of 145 excerpts of polyphonic music shows that the algorithms can efficiently classify drum components and compare favorably to a classic “bag-of-features” approach using support vector machines and spectral mid-level features.
ER -