Academy Research Fellow, Docent, D.Sc.(Tech)
Leader of the noise robust speech recognition team
Room G436 in Electrical Engineering building,
Otakaari 5, Otaniemi campus area, Espoo
- Mail address:
Aalto University School of Electrical Engineering,
Department of Signal Processing and Acoustics,
P.O. Box 13000, FI-00076 Aalto, Finland
kalle.palomaki (at) aalto.fi
- +358 50 430 1491
My main research interest is in noise robust automatic speech
recognition. Particularly I am interested in models that are inspired by
human auditory system, missing feature methods, binaural models and
methods based on sparse representations. One of the specific application
areas has been reverberation robust speech recognition. Other recent interests include audio based retrieval and creativity. During my doctoral studies
I have also conducted auditory brain research applying magnetoenchephalography. In the future I would like to build speech recognizers that are closer to human brains in neural networks and functions.
Noise robust speech recognition team
I am currently leader of Noise Robust Speech Recognition team and also a member of Speech Recognition Group
. We are also members of the National Centre of Excellence in Computational Inference
. My team members and alumni of the team are listed below.
- Ulpu Remes: missing feature imputation and uncertainty modeling in noise robust ASR, man to machine comparisons, noise robust speech synthesis.
- Sami Keronen: deep learning and missing feature mask estimation in noise robust ASR.
- Päivi Raino: Study of sign language and hearing impaired.
- Katri Leino: Medical dictation, speech based user interfaces, speaker adaptation.
- Heikki Kallasjoki: sparsity based methods and uncertainty modeling in noise robust ASR, reverberation robust ASR.
- Ana Ramirez-Lopez: man to machine comparisons, microphone arrays
- Annamaria Mesaros: audio content based retrieval, sound event recognition, singing recognition.
Selected recent publications:
Keronen S., Kallasjoki H., Palomäki K., Brown G., and
Gemmeke J. (2015) Feature enhancement of reverberant speech by distribution
matching and non-negative matrix factorization, EURASIP Journal on Advances in
Signal Processing, 76, eurasip
Remes U., Ramírez López A., Juvela L., Palomäki K., Brown G. J., Alku P., Kurimo M. (2016) Comparing human and automatic speech recognition in a perceptual restoration experiment, Computer Speech and Language, 35, 14-31, science-direct
Remes. U., Ramírez López A., Palomäki K. and Kurimo M. (2015) Bounded conditional mean imputation with observation uncertainties and acoustic model adaptation, IEEE Transactions on Audio Speech and Language Processing, 23(7),
Kallasjoki H., Gemmeke J. and Palomäki K. J. (2014) Estimating uncertainty to improve exemplar-based feature enhancement for noise robust speech recognition, IEEE Transactions on Audio Speech and Language Processing, 22(2), 368-380. ieee-xplore
Mesaros A., Heittola T. and Palomäki K. (2013) Query-by-example retrieval of sounds events using an integrated similarity measure of content and label, Proc. of The 14th International Workshop on Image and Audio Analysis for Multimedia Interactive Services (WIAMIS 2013). ieee-xplore, demo
Keronen S., Kallasjoki H., Remes U., Brown, G. J., Gemmeke J. F., and
Palomäki K. J. (2013) Mask estimation and imputation methods for missing
data speech recognition in a multisource reverberant environment,
Computer Speech and Language, 27(3), 798-819. science-direct
- Palomäki K. J. and Brown G. J. (2011) A computational model of
binaural speech recognition: role of across-frequency vs.
within-frequency processing and internal noise, Speech Communication,
53(6), 924-940. science-direct.
For a full list, see my publications page
. For citations see my Google Scholar
Software and demos
Links to web pages of some recent collaborators and collaboration themes.
- Guy Brown: Auditory inspired models, binaural models and reverberant speech recognition
- Paavo Alku: Speech enhancement, noise robust features, and formerly MEG brain measurements
- Jon Barker: Recognition of reverberant speech
- Jort Gemmeke: Sparse representations and uncertainty modeling
- Tuomas Virtanen: Sparse separation and noise models
- Amy Beeston: Auditory system inspired models in reverberant speech recognition
- Jouni Pohjalainen: Noise robust feature extraction
- Intelligent information acces to generic audio (2012-2013)
collaboration grant awarded by Center of Excellence in Computational
Inference, COIN .
- Academy Research Fellowship (2010-2015) on "Noise Robust Automatic
Speech Recognition: Searching for Answers from Human Hearing" the Academy of Finland. The fellowship is supplemented by a grant for establishment of a research team in noise robust ASR.
- Post doctoral researcher project (2007-2009) on "Auditory Approaches to Automatic speech recognition" by the Academy of Finland. The project included a grant for empoying a doctoral student.