BUSIM / SPG: Speech Processing Group

Lab Image: 

The speech group at Boğaziçi University was initiated in year 1998 as Dr. Arslan joined the department from Entropic Research (later acquired by Microsoft). In year 2005, Dr. Murat Saraclar joined the group from AT&T Labs Research. Currently there are 6 PhD and 2 MSc  students actively working on speech processing research. Main topics  include large vocabulary continuous speech recognition, voice conversion, 3-D audio-visual synthesis, speech synthesis, speech therapy and speaker identification. Until now 2 PhD dissertations and 17 MSc theses have been completed by the group members. The courses related to speech research at the department are as follows:

 Speech Processing, Psychoacoustics, Automatic Speech Recognition,  Advanced Speech Processing, Statistical Natural Language Processing.
 The group is involved in two European Union Network of Excellence  projects, namely SIMILAR and BIOSECURE. Aside from the cooperation  with the partners of those projects, there is active collaboration  with Boğaziçi University Linguistics Department, Istanbul Cerrahi  Hospital, and Anadolu University Speech Therapy Center. The spin-off
speech research company from the group, SESTEK (www.sestek.com.tr),  has products based on speech recognition, speech synthesis and voice  conversion technology.


Research Fields: 
Research Title: 
Voice Conversion
Research SubTitles: 
Cross-lingual voice conversion for music applications, conversion of pathological voices
Research Title: 
Broadcast News Transcription & Retrieval
Research SubTitles: 
Language modeling for Automatic Turkish Broadcast News Transcription, Spoken Term Detection
Research Title: 
Speaker Recognition
Research SubTitles: 
Robust speaker recognition for mismatch channel and handset conditions, NIST Speaker Recognition Evaluations
Laboratory People-Academics: 
Lab Page: