Title :
Unsupervised Speaker Change Detection for Mobile Device Recorded Speech
Author :
Vuorinen, O. ; Peltola, Jukka ; Makela, S. -M.
Author_Institution :
VTT Tech. Res. Centre of Finland, Oulu, Finland
Abstract :
In this paper we propose an unsupervised speaker change detection (SCD) system developed for mobile device applications. We use Bayesian information criterion (BIC) to find initial speaker changes, which are then verified or discarded in the second phase by utilizing modified BIC and silence detector information. Silence information usage after initial BIC in decision making is useful to separate real changes from noise peaks. Enhanced peak detector adjusts BIC penalty parameter automatically, which improve the robustness and feasibility. Improved BIC based false alarm compensation (FAC) merges effectively consecutive segments belonging to same speaker. Our experiments have shown the robustness of the algorithm and it produces very satisfactory results for difficult mobile phone recorded speech data.
Keywords :
Bayes methods; decision making; mobile radio; speaker recognition; speech processing; Bayesian information criterion; decision making; false alarm compensation; mobile device speech recording; mobile phone; unsupervised speaker change detection; Audio recording; Bayesian methods; Change detection algorithms; Detectors; Image segmentation; Indexing; Multimedia databases; Noise robustness; Speech analysis; Video recording; metadata; mobile audio segmentation; multimedia database; speaker change detection; speaker segmentation;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0727-3
DOI :
10.1109/ICASSP.2007.366346