Erfan Loweimi

Contact Information
Mobile: +44 (0)7553 261 989 Phone: +44 (0)114 222 1800 Fax: +44 (0)114 222 1810.
G043, Regent Court, 211 Portobello, Sheffield, UK, S1 4DP
Erfan Loweimi
I completed my Ph.D. in February 2018 in the Department of Computer Science at the University of Sheffield under the supervision of two great scholars, Professor Jon Barker and Professor Thomas Hain in the SPandH and MINI groups. My PhD thesis title is “Robust Phase-based Speech Signal Processing; From Source-Filter Separation to Model-Based Robust ASR”.
My research interests revolve around phase-based speech signal processing and recognition. From signal processing point of view, I am interested in evaluating the information content of the phase spectrum along with the way which information is encoded in this data sequence. From speech recognition standpoint, I am interested in looking for novel representations and feature extraction methods from the phase spectrum which could render high discriminability and robustness with a tractable complexity. In speech synthesis, I am interested in the phase-only signal reconstruction and improving the quality/intelligibility of the phase-only reconstructed speech. Last but not least, statistical studies of the speech phase spectrum and its sensitivity to noise constitute other parts of my works [ICASSP 2017]. All these are done in a comparative manner and the same process is carried out for the magnitude spectrum which basically constitutes the baseline.
I am also interested in the general problem of speech recognition in noisy condition and has developed a technique called generalised-VTS [Interspeech16,Interspeech17]. It provides a remarkable accuracy improvement in the presence of both additive and channel noises. Although it is originally designed for robust ASR, it reduces the word error rate in the clean-matched condition, too. Extracting features capture the long-term properties of speech signal such as emotion is also of my interest [SLSP15].
For more details about my scientific background and publications do please look at my CV.
2017
-
E. Loweimi, J. Barker, and T. Hain, “Statistical Normalisation of Phase-based Feature Representation for Robust Speech Recognition,” in ICASSP 2017, New Orleans, USA, 2017.
[Bibtex]@inproceedings{loweimi_icassp17, address = {New Orleans, USA}, author = {Erfan Loweimi and Jon Barker and Thomas Hain}, booktitle = {{ICASSP 2017}}, title = {{Statistical Normalisation of Phase-based Feature Representation for Robust Speech Recognition}}, year = {2017} }
-
E. Loweimi, J. Barker, and T. Hain, “Channel Compensation in the Generalised Vector Taylor Series Approach to Robust ASR,” in Interspeech 2017, Stockholm, Sweden, 2017.
[Bibtex]@inproceedings{loweimi_is17-gvts, address = {Stockholm, Sweden}, author = {Erfan Loweimi and Jon Barker and Thomas Hain}, booktitle = {{Interspeech 2017}}, title = {{Channel Compensation in the Generalised Vector Taylor Series Approach to Robust ASR}}, year = {2017} }
-
E. Loweimi, J. Barker, O. S. Torralba, and T. Hain, “Robust Source-filter Separation of Speech Signal in the Phase Domain,” in Interspeech 2017, Stockholm, Sweden, 2017.
[Bibtex]@inproceedings{loweimi_is17-ph, address = {Stockholm, Sweden}, author = {Erfan Loweimi and Jon Barker and Oscar Saz Torralba and Thomas Hain}, booktitle = {{Interspeech 2017}}, title = {{Robust Source-filter Separation of Speech Signal in the Phase Domain}}, year = {2017} }
2016
-
E. Loweimi, J. Barker, and T. Hain, “Use of Generalised Nonlinearity in Vector Taylor Series Noise Compensation for Robust Speech Recognition,” in Proceedings of the 17th Annual Conference of the International Speech Communication Association (Interspeech), San Francisco, USA, 2016.
[Bibtex]@inproceedings{loweimi_is16, address = {San Francisco, USA}, author = {Erfan Loweimi and Jon Barker and Thomas Hain}, booktitle = {{Proceedings of the 17th Annual Conference of the International Speech Communication Association (Interspeech)}}, title = {{Use of Generalised Nonlinearity in Vector Taylor Series Noise Compensation for Robust Speech Recognition}}, year = {2016} }
2015
-
E. Loweimi, J. Barker, and T. Hain, “Source-filter Separation of Speech Signal in the Phase Domain,” in Proceedings of the 16th Annual Conference of the International Speech Communication Association (Interspeech), Dresden, Germany, 2015.
[Bibtex]@inproceedings{loweimi_is15, address = {Dresden, Germany}, author = {Erfan Loweimi and Jon Barker and Thomas Hain}, booktitle = {{Proceedings of the 16th Annual Conference of the International Speech Communication Association (Interspeech)}}, title = {{Source-filter Separation of Speech Signal in the Phase Domain}}, year = {2015} }
-
E. Loweimi, M. Doulaty, J. Barker, and T. Hain, “Long-term statistical Feature Extraction from Speech Signal and its Application in Emotion Recognition,” in Statistical Language and Speech Processing (SLSP), Budapest, Hungary, 2015.
[Bibtex]@inproceedings{loweimi_slsp15, address = {Budapest, Hungary}, author = {Erfan Loweimi and M. Doulaty and Jon Barker and Thomas Hain}, booktitle = {{Statistical Language and Speech Processing (SLSP)}}, title = {{Long-term statistical Feature Extraction from Speech Signal and its Application in Emotion Recognition}}, year = {2015} }
-
E. Loweimi, J. Barker, and T. Hain, “Emotion Recognition from Speech Signal by Effective Combination of the Generative and Discriminative Models,” in University of Sheffield Engineering Symposium, Sheffield, UK, 2015.
[Bibtex]@inproceedings{loweimi_uses15, address = {Sheffield, UK}, author = {Erfan Loweimi and Jon Barker and Thomas Hain}, booktitle = {{University of Sheffield Engineering Symposium}}, title = {{Emotion Recognition from Speech Signal by Effective Combination of the Generative and Discriminative Models}}, year = {2015} }
2014
-
E. Loweimi, J. Barker, and T.Hain, “Compression of model-based group delay function for robust speech recognition,” in University of Sheffield Engineering Symposium, Sheffield, UK, 2014.
[Bibtex]@inproceedings{loweimi_uses14, address = {Sheffield, UK}, author = {Erfan Loweimi and Jon Barker and T.Hain}, booktitle = {{University of Sheffield Engineering Symposium}}, title = {{Compression of model-based group delay function for robust speech recognition}}, year = {2014} }
2013
-
E. Loweimi, S. M. Ahadi, and T. Drugman, “A new phase-based feature representation for robust speech recognition,” in IEEE International conference on Acoustics, Speech and Signal Processing, Vancouver, Canada, 2013.
[Bibtex]@inproceedings{loweimi_icassp13, address = {Vancouver, Canada}, author = {Erfan Loweimi and Seyed Mohammad Ahadi and Thomas Drugman}, booktitle = {{IEEE International conference on Acoustics, Speech and Signal Processing}}, title = {{A new phase-based feature representation for robust speech recognition}}, year = {2013} }
-
E. Loweimi, S. M. Ahadi, T. Drugman, and S. Loveymi, “On the importance of pre-emphasis and window shape in phase-based speech recognition,” in Lecture Notes in Computer Science, Advances in Non-Linear Speech Processing (NOLISP), Mons, Belgium, 2013.
[Bibtex]@inproceedings{loweimi_nolisp13, address = {Mons, Belgium}, author = {Erfan Loweimi and Seyed Mohammad Ahadi and Thomas Drugman and Samira Loveymi}, booktitle = {{Lecture Notes in Computer Science, Advances in Non-Linear Speech Processing (NOLISP)}}, title = {{On the importance of pre-emphasis and window shape in phase-based speech recognition}}, year = {2013} }
2011
-
E. Loweimi, S. M. Ahadi, and S. Loveymi, “On the importance of phase and magnitude spectra in speech enhancement,” in Iranian conference on Electrical Engineering (ICEE), Tehran, Iran, 2011.
[Bibtex]@inproceedings{loweimi_icee11, address = {Tehran, Iran}, author = {Erfan Loweimi and Seyed Mohammad Ahadi and Samira Loveymi}, booktitle = {{Iranian conference on Electrical Engineering (ICEE)}}, title = {{On the importance of phase and magnitude spectra in speech enhancement}}, year = {2011} }
-
E. Loweimi and S. M. Ahadi, “A new group delay-based feature for robust speech recognition,” in IEEE International conference on Multimedia and Expo (ICME), Barcelona, Spain, 2011.
[Bibtex]@inproceedings{loweimi_icme11, address = {Barcelona, Spain}, author = {Erfan Loweimi and Seyed Mohammad Ahadi}, booktitle = {{IEEE International conference on Multimedia and Expo (ICME)}}, title = {{A new group delay-based feature for robust speech recognition}}, year = {2011} }
-
E. Loweimi, S. M. Ahadi, and H. Sheikhzadeh, “Phase-only speech reconstruction using very short frames,” in Proceedings of the 12th Annual Conference of the International Speech Communication Association (Interspeech), Florence, Italy, 2011.
[Bibtex]@inproceedings{loweimi_is11, address = {Florence, Italy}, author = {Erfan Loweimi and Seyed Mohammad Ahadi and Hamid Sheikhzadeh}, booktitle = {{Proceedings of the 12th Annual Conference of the International Speech Communication Association (Interspeech)}}, title = {{Phase-only speech reconstruction using very short frames}}, year = {2011} }
2010
-
E. Loweimi and S. M. Ahadi, “Objective evaluation of magnitude and phase only spectrum-based reconstruction of the speech signal,” in International Symposium on Communications, Control and Signal Processing, Limassol, Cyprus, 2010.
[Bibtex]@inproceedings{loweimi_isccsp10, address = {Limassol, Cyprus}, author = {Erfan Loweimi and Seyed Mohammad Ahadi}, booktitle = {{International Symposium on Communications, Control and Signal Processing}}, title = {{Objective evaluation of magnitude and phase only spectrum-based reconstruction of the speech signal}}, year = {2010} }
-
E. Loweimi and S. M. Ahadi, “Objective evaluation of phase and magnitude only reconstructed speech: new considerations,” in Information Sciences Signal Processing and their Applications (ISSPA), Kuala Lumpur, Malaysia, 2010.
[Bibtex]@inproceedings{loweimi_isspa10, address = {Kuala Lumpur, Malaysia}, author = {Erfan Loweimi and Seyed Mohammad Ahadi}, booktitle = {{Information Sciences Signal Processing and their Applications (ISSPA)}}, title = {{Objective evaluation of phase and magnitude only reconstructed speech: new considerations}}, year = {2010} }
My viva took place on January 15, 2018, and was successfully passed with minor amendments. I had two great scholars as examiners: Professor Yannis Stylianou and Professor Guy Brown.
The eThesis “Robust Phase-based Speech Signal Processing; From Source-Filter Separation to Model-Based Robust ASR” is now available at the following URL: http://etheses.whiterose.ac.
Channel Compensation in the Generalised Vector Taylor Series to Robust Automatic Speech Recognition
— UK Speech 2017, Cambridge , UK. 11, September, 2017 Slides Video
— Interspeech 2017, Stockholm , Sweden. 23, August, 2017
Robust Automatic Speech Recognition using Generalised Vector Taylor Series
— Engineering Research Symposium, Sheffield, UK. 30, June, 2017
A Genie in the mike! The Science of Talking (with) Machines
— A Pint of Science Festival, Sheffield, UK. 15, May, 2017 Teaser
Signal Processing is Dead (?)! Long Live DNN!
— Machine Intelligence for Natural Interfaces (MINI) workshop, Sheffield, UK. 12, December, 2016
Robust Automatic Speech Recognition: A Whirlwind Tour
— Acoustics Exchange Day, Sheffield, UK. 13, September, 2016
Use of Generalised Non-linearity in Vector Taylor Series Noise Compensation for Robust Speech Recognition
— SPandH Seminar, Sheffield, UK. 3, August, 2016
Deep Learning, The End of History and The Last Computer Scientist
— Sheffield Robotics Seminar, Sheffield, UK. 5, July, 2016
Deep Learning – The Last Computer Scientist
— A Pint of Science Festival, Sheffield, UK. 24, May, 2016
MINI Feature Extraction Toolkit for Automatic Speech Recognition
— Machine Intelligence for Natural Interfaces (MINI) workshop, Sheffield, UK. 1, December, 2015
Long-term Statistical Feature Extraction from Speech Signal and its Application in Emotion Recognition
— Statistical Language and Speech Processing (SLSP), Budapest, Hungary. 25, November, 2015
Source-filter Separation of Speech Signal in the Phase Domain
— UKSpeech 2015, Norwich, UK. 3, July, 2015
Why PhD?
— Department of Computer Science, University of Sheffield, Sheffield, UK. 26, November, 2014
On the Importance of Phase Spectrum in Speech Signal Processing
— SPandH Seminar, Sheffield, UK. 11, November, 2014
Compression of the Model-based Group Delay Function for Robust Speech Recognition
— University of Sheffield Engineering Symposium (USES), Sheffield, UK. 24, July, 2014
Ethics in Data Modelling; Love it or Leave it?
— Research Ethics and Integrity module, Engineering Faculty, University of Sheffield, Sheffield, UK. 4, March, 2014
Research Communicator of the Year Award, Faculty of Engineering, University of Sheffield. 30, June, 2017
ISCA Student Travel Grant, Interspeech 2017, Stockholm , Sweden. August, 2017
Faculty PhD Scholarship, Faculty of Engineering, Department of Computer Science, University of Sheffield. 1, October, 2013
I received my B.Sc. and M.Sc. degrees in Electronics Engineering from the Shahid Chamran University of Ahvaz and Amirkabir University of Technology (Tehran Polytechnic), in 2007 and 2010, respectively. My B.Sc. final project was about Software-defined Radio and I did it under the supervision of Dr. Hooman Kaabi. My Master project was titled On the Importance of Phase in Robust Speech Recognition. During my M.Sc. project I enjoyed working with Professor Seyed Mohammad Ahadi as the supervisor and Dr. Hamid Sheikhzadeh as the adviser.
I am an avid book reader, mainly history, politics and stories. In the history genre, I am essentially interested in two periods, history of the Early Islam (the first 61 years) and Iran’s contemporary history (From Iranian Constitutional Revolution onwards). In politics, I closely follow the Middle-East events and monitor the geopolitical trends on that volatile part of the globe. I also enjoy sports and regularly exercise, especially swimming and like the martial arts. My favourite fighters are Fedor Emelianenko, Mirko Filipović (Cro Cop), Andy Hug, Francisco Filho and Kenji Midori.
Being an entrepreneur is among my ultimate goals in life. In my spare time (if luckily find any!) I watch/read about the success stories of the great people (among others, Steve Jobs is my favourite) and try to learn and compile the attributes which make the inspiring people inspiring. The big secret is that there is no secret and living up to the dreams is absolutely possible if …