Title :
Extracting the prosodic information for Turkish broadcast news data and using on the sentence segmentation task
Author :
Dalva, Dogan ; Revidi, Izel D. ; Guz, Umit ; Gurkan, Hakan
Author_Institution :
Elektrik-Elektron. Muhendisligi Bolumu, Isik Univ., Istanbul, Turkey
Abstract :
In this study, extracting the prosodic information for Turkish Broadcast News Data using the open source tools and comparing the sentence segmentation performances of these grouped prosodic information on the raw data obtained as an output from the Automatic Speech Recognition System are established. Especially for the sentence segmentation task, a very promising prosodic feature set is obtained.
Keywords :
feature extraction; natural language processing; public domain software; speech recognition; Turkish broadcast news data; automatic speech recognition system; open source tools; prosodic feature set; prosodic information extraction; sentence segmentation task; Conferences; Entropy; Hidden Markov models; NIST; Signal processing; Speech; Training data; automatic speech segmentation; prosodic feature set; prosody; sentence segmentation;
Conference_Titel :
Signal Processing and Communications Applications Conference (SIU), 2014 22nd
Conference_Location :
Trabzon
DOI :
10.1109/SIU.2014.6830603
