8th International Conference on Spoken Language Processing

Jeju Island, Korea
October 4-8, 2004

Learning Nonnegative Features of Spectro-Temporal Sounds for Classification

Yong-Choon Cho (1), Seungjin Choi (2)

(1) Samsung, Korea
(2) POSTECH, Korea

In this paper we present a method of sound classification which exploits a parts-based representation of spectro-temporal sounds, employing the nonnegative matrix factorization (NMF). We illustrate a new way of learning nonnegative features using a variant of NMF and show its useful behavior in the task of general sound classification with comparison to independent component analysis (ICA) which produces holistic features.

Full Paper

Bibliographic reference.  Cho, Yong-Choon / Choi, Seungjin (2004): "Learning nonnegative features of spectro-temporal sounds for classification", In INTERSPEECH-2004, 989-992.