Hossein Sameti

 

Hossein Sameti
Assistant Professor

Computer Engineering Department
Sharif University of Technology

  Tel

+98 21 66164637

  Fax

+98 21 66019246

  Mailing add.

Azadi Av. Sharif University of Technology, Computer Engineering Dep., Room No. 306

  E-Mail

sameti [at] sharif [dot] edu

 

About me

Education

Research Interests

 Publications

Courses

Students

Speech Processing Lab

 

 :: About me

 

 

Director of the Speech Processing Lab (SPL), Sharif University of Technology

This research group started to work in 1999. It is working towards preparing a Persian (Farsi) speech recognition engine which can be utilized in different speech recognition products. The project includes research in different areas of automatic speech recognition and incorporation of different known methods for Persian (Farsi0 language. The system is a Large Vocabulary Speaker-Independent Continuous Speech Recognition (LV-SICSR) system based on Hidden Markov Models (HMM) with a selectable lexicon of at least 1000 words. The first versions of this system so-called NEVISA, are completed which includes many robustness methods and NLP approaches. The improvements of this system for 10,000 words is in progress now. Presently, 25 researchers are involved in the project. For more information about the group see spl.ce.sharif.edu. Also our official web site, www.asr-gooyesh.com contains more detailed information about our speech products.

 

 

 

Up

 :: Education

 

 

 

 

 

Up

 :: Research Interests

 

 
  • Speech Processing
  • Automatic Speech Recognition
  • Hidden Markov Modeling of Speech
  • Speech Enhancement

 :: Publications

 

 
  1. Momtazi S., Fazel M., Sameti H., Bahrani M.,  Robust Parsing for Word Lattices in Continuous Speech Recognition Systems , 11th International CSI Computer Conference, Tehran, Iran, Jan 2006. (in Persian) 
  2. Bahrani M., Sameti H.,  Extraction and Modeling Context Dependent Phone Units for Improvement of Continuous Speech Recognition Accuracy by Phonemes Clustering , Submitted in Iranian Journal of Electrical and Computer Engineering, Vol 2, 2005. (in Persian) 
  3. Hafezi N., Bahrani M., Movasagh H., Sameti H.,  Extracting Statistical Language Models for Continuous Speech Recognition Systems Using Persian Text Corpus , 7th Conference of Intelligent Systems, Dec 2005. (in Persian)
  4. Sajedi H., Babaali B., Sameti H., Incorporating the Genetic algorithm techniques in the training process of a Speech Recognition System, the 14th Iranian Conference in Electrical Engineering (ICEE), Tehran-Iran, 2005. 
  5. Halavati R., Bagheri Shouraki S., Sameti H., Harati Zadeh S., Babaali B.,  A Novel Noise Immune Speech Recognition Approach , 11th IFSA World Congress, Beijing, China, 2005. - BEST STUDENT PAPER  
  6. Babaali B., Bagheri M.R, Hosseinzadeh Kh., Bahrani M., sameti H., A phoneme to word decoder based on lexicon tree for Persian speech recognition, 10th Annual Computer Society of Iran Computer Conference (CSICC), Tehran-Iran, 2005.   
  7. Bahrani M., Sameti H., Robust Parsing for Word Lattices in Continuous Speech Recognition Systems, ISSPIT 2005.
  8. Movasagh H.,Bahrani M., Sameti H.,  Building and Incorporating Language Models for Persian continuous speech Recognition system , ISSPIT 2005.
  9. Movasagh H., Sameti H., Two–step synchronous phoneme based search for continuose speech Recognition, ICM SAO, 2005.
  10. Veisi H., Fazel A., Sameti H.,  Increasing the digit recognition rate in noise using noise robust features , Iranian Conference in Electrical Engineering (ICEE),  Zanjan, Iran, 2005.
  11. Veisi H., Sameti H., Abutalebi H.R.,  On the using of Principal Component Analysis on the Farsi Continuous Speech Recognition for Feature Robustness and Feature Reductions , 10th Annual Computer Society of Iran Computer Conference (CSICC), Tehran-Iran, 2005. (in Persian)
  12. Fazel Dehkordi A., Sameti H., Bahrani M.,  Designing and Preparing of FARSI Connected Digit Database and Using it in a Word Based Automatic Speech Recognition System , 13th Iranian Conference on Electrical Engineering (ICEE), Zanjan, Iran, 2005. (in Persian)
  13. Fazel Dehkordi A., Sameti H., Manzuri M. T., RCC-Mean Subtraction Robust Feature and Compare Various Feature based Methods for Robust Speech Recognition in presence of Telephone Noise, 10th Annual Conference of Computer Society of Iran, Tehran, Iran, 2005. (in Persian)
  14. Fazel Dehkordi A., Sameti H., Ghiathi S. K,  A new feature extraction motivated by human , First International Conference On Modeling, Simulation and Applied Optimization,  (ICMSAO’05), Sharjah, UAE, 2005.
  15. Hosseinzadeh K., Sameti H., Abutalebi H.R. , Fazel Dehkordi A.,  MLLR method for environmental adaptation in the continuous FARSI speech recognition , The 6th Conference on Intelligent Systems, Kerman, Iran, 2004.
  16. Babaali B., Sameti H.,  The Sharif Speaker-Independent Large Vocabulary Speech Recognition System , The 2nd Workshop on Information Technology & Its Disciplines (WITID 2004), Feb. 24-26, 2004, Kish Island, Iran.
  17. Sameti H., Movasagh H., Babaali B., Bahrani M., Hosseinzadeh K., Fazel Dehkordi A., Abutalebi H. R., Veisi H., Mokri Y., Motazeri N., Nezami Ranjbar M.,  Large Vocabulary Persian Speech Recognition System , The 1st workshop on Persian language and computer, Tehran, Iran,  May 25-26 2004. (in Persian)
  18. H. Sameti and Li Deng, Nonstationary-State Hidden Markov Model Representation of Speech Signals for Speech Enhancement, Elsevier Signal Processing Journal, Volume 82, Number 2, pp. 205-227, Feb. 2002.
  19. H. Sameti, H. Sheikhzadeh, L. Deng, and R. L. Brennan, HMM-Based Strategies for Enhancement of Speech Embedded in Non-Stationary Noise , IEEE Trans. on Speech and Audio Processing, Vol. 6, No. 5, pp. 445-455, Sept. 1998.
  20. H. Sameti, Nonstationary - State Hidden Markov Models for Speech Enhancement , The Second Iranian Computer Society Conference, Amirkabir University of  Technology, Tehran, Iran, December 1996.
  21. L. Deng and H. Sameti, Transitional Speech Units and Their Representation by Regressive Markov States: Applications to Speech Recognition, IEEE Trans. on Speech and Audio Processing, Vol. 4, No. 4, pp. 301-306, July 1996.
  22. L. Deng, J. Wu, and H. Sameti, Improved Speech Modeling and Recognition Using Multi-dimensional Articulatory States as Primitive Speech Units, Proceedings of IEEE Int. Conf. Acoustics, Speech, and Signal Processing (ICASSP), pp. 385-388, Detroit, USA, April 1995.
  23. H. Sheikhzadeh, R. L. Brennan, and H. Sameti, Real-Time Implementation of HMM-Based MMSE Algorithm for Speech Enhancement in Hearing Aid Applications, Proceedings of IEEE Int. Conf. Acoustics, Speech, and Signal Processing (ICASSP), pp. 808-811, Detroit, USA, April 1995.
  24. L. Deng and H. Sameti, Automatic Speech Recognition Using Dynamically Defined Speech Units , Proceedings of the 1994 International Conference on Spoken Language Processing, Vol. 4, pp. 2167--2170, Yokohama, Japan, September 18-22, 1994.
  25. L. Deng and H. Sameti, Articulatory Phonology and Speech Recognition: A Case Study Involving Use of Dynamically Defined Speech Primitives, Acoustical Society of America, 127th Meeting, Cambridge, MA, June 1994.
  26. H. Sheikhzadeh, H. Sameti, L. Deng, and R. L. Brennan, Comparative Performance of Spectral Subtraction and HMM-Based Speech Enhancement Strategies with Application to Hearing Aid Design , Proceedings of IEEE Int. Conf. Acoustics, Speech, and Signal Processing (ICASSP), pp. I-13, I-16, Adelaide, Australia, April 1994.

 

 

 

 :: Courses

 

 
  •  Speech Recognition, Graduate
  •  Speech Processing,  Graduate
  •  Speech Enhancement,  Graduate
  •  Statistics

 

 

 

 :: Students

 

 

Ph.D. Students

 

 

 

 1. Babaali Bagher, Starting 2004
 2. Bahrani Mohammad,  Starting 2004

 

 

 

 

 

M.S. Students

 

 

 

 1. Veisipour Saman, Statring 2004
    Thesis: Confidence measure and out of vocabulary

 

 

 

 2. Safayani Mehran,  Statring 2004
    Thesis: Using microphone array for improving speech recognition performance

 

 

 

 3. Veisi Hadi, Nov. 2005
    Thesis: Model based methods for robust speech recognition systems

 

 

 

 4. Hosseinzadeh Khosro, Nov. 2004
    Thesis: Improving the accuracy of speech recognition in  noisy environments

 

 

 

 5. Movasagh Hamed, Jan. 2003
    Thesis: Design and Implementation of Optimized Search for Persian Speech Recognition by HMM

 

 

 

 :: Speech Processing Lab (SPL)

 

 

SPL is one of the research laboratories of Computer Engineering Department of Sharif University Of Technology. The main area of the research in this Lab is Digital Signal Processing, specially Speech signals. Speech Recognition is the major field of activity for this group and a speech recognition engine including the latest and successful algorithms is developed. This engine is customized for Persian language including Persian language models (Statistical and Grammatical) and the first version is released as NEVISA, the most robust and accurate Persian dictation system. SPL also developed NEWSHA, the high accurate Persian telephony speech recognition including command and digit recognition. The summary of the fields of activity of this group are:

  • Digital Speech Processing
  • Robust Speech Recognition
  • Speech Synthesis (Text-To-Speech)
  • Speech Enhancement
  • Natural Language Processing (NLP)
  • Speech Database Design
  • Pattern Recognition

More information about this group is available in spl.ce.sharif.edu page.

 

 

 

 Speech Processing 86-2 Grades