Abstract | ||
---|---|---|
We propose a novel principle based on Complex Spec- trum Circle Centroid (CSCC) for restoring complex spec- trum of the target signal from multiple microphone input signals in a noisy environment. If noise arrives at mul- tiple microphones with different time delays relative to the target signal, the observed noisy signals lie on a cir- cle in the complex spectrum plane from which the tar- get signal is restored by finding the centroid of the cir- cle. Unlike most of existing methods for noise reduc- tion such as ICA, AMNOR and beamforming, this non- linear operation is applicable to any type of noise in- cluding non-stationary, moving, signal-correlated, non- planar, and spoken noises, without identifying the noise direction and training parameters. The proposed method was evaluated with speech recognition experiments in simulated noisy environments and was shown to improve the word accuracy close to the clean speech recognition rate of 89.4% in the case of a single spoken noise, and from 0% with one micro- phone to 60.6% with 8 microphones in the case of 3 spo- ken noises. The properties of this new method is further discussed theoretically and experimentally. |
Year | Venue | Keywords |
---|---|---|
2004 | INTERSPEECH | linear operator,spectrum,speech recognition |
Field | DocType | Citations |
Noise reduction,Beamforming,Speech processing,Nonlinear system,Computer science,Speech recognition,Microphone array,Microphone,Centroid,Word accuracy | Conference | 1 |
PageRank | References | Authors |
0.38 | 4 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shigeki Sagayama | 1 | 1217 | 137.97 |
Okajima Takashi | 2 | 1 | 0.38 |
Kamamoto Yutaka | 3 | 1 | 0.72 |
Takuya Nishimoto | 4 | 227 | 28.95 |