Source separation using dictionary learning and deep recurrent neural network with locality preserving constraint

Tuan Pham, Yuan Shan Lee, Seksan Mathulaprangsan, Jia Ching Wang

研究成果: 書貢獻/報告類型會議論文篇章同行評審

3 引文 斯高帕斯(Scopus)

摘要

Deep learning is a popular method for monaural source separation, and especially for extracting a singing voice from a single-channel song. However, deep learning-based source separation ignores the geometrical structure of the input data. This work develops a novel approach to source separation that is based on non-negative matrix factorization (NMF) and deep recurrent neural networks (DRNN) with a locality-preserving constraint. First, NMF was used to learn patterns from training data. The learned patterns are linearly combined with the output of DRNN. Second, a locality-preserving constraint is developed to exploit the inner-structure of the input data in the DRNN learning process. Experimental results obtained using the MIR-1K dataset reveal that the proposed algorithm outperforms the baselines.

原文???core.languages.en_GB???
主出版物標題2017 IEEE International Conference on Multimedia and Expo, ICME 2017
發行者IEEE Computer Society
頁面151-156
頁數6
ISBN(電子)9781509060672
DOIs
出版狀態已出版 - 28 8月 2017
事件2017 IEEE International Conference on Multimedia and Expo, ICME 2017 - Hong Kong, Hong Kong
持續時間: 10 7月 201714 7月 2017

出版系列

名字Proceedings - IEEE International Conference on Multimedia and Expo
ISSN(列印)1945-7871
ISSN(電子)1945-788X

???event.eventtypes.event.conference???

???event.eventtypes.event.conference???2017 IEEE International Conference on Multimedia and Expo, ICME 2017
國家/地區Hong Kong
城市Hong Kong
期間10/07/1714/07/17

指紋

深入研究「Source separation using dictionary learning and deep recurrent neural network with locality preserving constraint」主題。共同形成了獨特的指紋。

引用此