Hirokazu Kameoka mainly investigates Speech recognition, Algorithm, Spectrogram, Non-negative matrix factorization and Artificial intelligence. In the subject of general Speech recognition, his work in Speech synthesis is often linked to Generator, thereby combining diverse domains of study. The various areas that Hirokazu Kameoka examines in his Algorithm study include Cluster analysis, Statistical model, Rule-based machine translation, Mixture model and Convolutional neural network.
His Non-negative matrix factorization study is associated with Matrix decomposition. His studies link Pattern recognition with Artificial intelligence. His Pattern recognition research incorporates themes from Iterative method, Source separation and Speech processing.
His scientific interests lie mostly in Speech recognition, Artificial intelligence, Pattern recognition, Algorithm and Spectrogram. His Speech recognition research is multidisciplinary, incorporating perspectives in Sequence and Fundamental frequency. Statistical model is closely connected to Natural language processing in his research, which is encompassed under the umbrella topic of Artificial intelligence.
His work on Feature extraction, Mixture model and Semi-supervised learning as part of general Pattern recognition study is frequently linked to Sparse matrix, therefore connecting diverse disciplines of science. Hirokazu Kameoka has researched Algorithm in several fields, including Frequency domain, Fourier transform, Spectral envelope and Linear predictive coding. While the research belongs to areas of Spectrogram, he spends his time largely on the problem of Source separation, intersecting his research to questions surrounding Autoencoder.
His primary areas of study are Speech recognition, Spectrogram, Algorithm, Naturalness and Source separation. His work on Sound quality as part of general Speech recognition research is frequently linked to Generator, bridging the gap between disciplines. His studies in Spectrogram integrate themes in fields like Speech enhancement, Wiener filter and Cluster analysis.
Hirokazu Kameoka focuses mostly in the field of Algorithm, narrowing it down to matters related to Waveform and, in some cases, Aliasing and Spectral amplitude. His study on Source separation is intertwined with other disciplines of science such as Matrix decomposition and Non-negative matrix factorization. Hirokazu Kameoka combines subjects such as Classifier, Pattern recognition and Generative model with his study of Autoencoder.
The scientist’s investigation covers issues in Speech recognition, Autoencoder, Speech synthesis, Source separation and Artificial intelligence. His study of Sound quality is a part of Speech recognition. His Source separation study integrates concerns from other disciplines, such as Spectrogram and Pattern recognition.
Hirokazu Kameoka integrates Spectrogram with Matrix decomposition in his research. Hirokazu Kameoka studies Mixture model which is a part of Artificial intelligence. As a part of the same scientific study, Hirokazu Kameoka usually deals with the Separation, concentrating on Blind signal separation and frequently concerns with Algorithm.
This overview was generated by a machine learning system which analysed the scientist’s body of work. If you have any feedback, you can contact us here.
Multichannel Extensions of Non-Negative Matrix Factorization With Complex-Valued Data
H. Sawada;H. Kameoka;S. Araki;N. Ueda.
IEEE Transactions on Audio, Speech, and Language Processing (2013)
StarGAN-VC: non-parallel many-to-many Voice Conversion Using Star Generative Adversarial Networks
Hirokazu Kameoka;Takuhiro Kaneko;Kou Tanaka;Nobukatsu Hojo.
spoken language technology workshop (2018)
Determined blind source separation unifying independent vector analysis and nonnegative matrix factorization
Daichi Kitamura;Nobutaka Ono;Hiroshi Sawada;Hirokazu Kameoka.
IEEE Transactions on Audio, Speech, and Language Processing (2016)
A Multipitch Analyzer Based on Harmonic Temporal Structured Clustering
H. Kameoka;T. Nishimoto;S. Sagayama.
IEEE Transactions on Audio, Speech, and Language Processing (2007)
Complex NMF: A new sparse representation for acoustic signals
Hirokazu Kameoka;Nobutaka Ono;Kunio Kashino;Shigeki Sagayama.
international conference on acoustics, speech, and signal processing (2009)
Separation of a monaural audio signal into harmonic/percussive components by complementary diffusion on spectrogram
Nobutaka Ono;Kenichi Miyamoto;Jonathan Le Roux;Hirokazu Kameoka.
european signal processing conference (2008)
Parallel-Data-Free Voice Conversion Using Cycle-Consistent Adversarial Networks
Takuhiro Kaneko;Hirokazu Kameoka.
arXiv: Machine Learning (2017)
Convergence-guaranteed multiplicative algorithms for nonnegative matrix factorization with β-divergence
Masahiro Nakano;Hirokazu Kameoka;Jonathan Le Roux;Yu Kitano.
international workshop on machine learning for signal processing (2010)
A REAL-TIME EQUALIZER OF HARMONIC AND PERCUSSIVE COMPONENTS IN MUSIC SIGNALS
Nobutaka Ono;Kenichi Miyamoto;Hirokazu Kameoka;Shigeki Sagayama.
international symposium/conference on music information retrieval (2008)
CycleGAN-VC: Non-parallel Voice Conversion Using Cycle-Consistent Adversarial Networks
Takuhiro Kaneko;Hirokazu Kameoka.
european signal processing conference (2018)
If you think any of the details on this page are incorrect, let us know.
We appreciate your kind effort to assist us to improve this page, it would be helpful providing us with as much detail as possible in the text box below: