Name
Title | ||
|---|---|---|
Phase reconstruction from amplitude spectrograms based on directional-statistics deep neural networks |
Abstract | ||
|---|---|---|
•We propose phase reconstruction methods from amplitude spectrograms using directional statistics deep neural networks (DNNs).•The directional statistics DNN is a novel deep generative model that has a circular probability distribution as the conditional probability.•We use the DNN to model not only phase of speech signals but also group delay that is strongly related to amplitude spectra.•Experimental evaluation demonstrates that our method outperforms the conventional signal processing based method. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1016/j.sigpro.2019.107368 | Signal Processing |
Keywords | Field | DocType |
Speech analysis,Phase reconstruction,Deep neural network,Directional statistics,Group delay | Speech processing,Directional statistics,Histogram,Pattern recognition,Control theory,Spectrogram,von Mises distribution,Group delay and phase delay,Artificial intelligence,Artificial neural network,Cardioid,Mathematics | Journal |
Volume | ISSN | Citations |
169 | 0165-1684 | 1 |
PageRank | References | Authors |
0.48 | 0 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Shinnosuke Takamichi | 1 | 75 | 22.08 |
| Saito, Yuki | 2 | 26 | 7.87 |
| Norihiro Takamune | 3 | 35 | 10.18 |
| Daichi Kitamura | 4 | 142 | 21.21 |
| Saruwatari, H. | 5 | 652 | 90.81 |