Loading...
Please wait, while we are loading the content...
Improved modeling of cross-decoder phone co-occurrences in SVM-based phonotactic language recognition (2010)
| Content Provider | CiteSeerX |
|---|---|
| Author | Penagarikano, Mikel Varona, Amparo Rodríguez-Fuentes, Luis Javier Bordel, Germán |
| Description | in Proc. Odyssey: the Speaker and Language Recognition Workshop Most common approaches to phonotactic language recognition deal with several independent phone decod-ings. These decodings are processed and scored in a fully uncoupled way, their time alignment (and the infor-mation that may be extracted from it) being completely lost. Recently, a new approach to phonotactic language recognition has been presented [1], which takes into ac-count time alignment information, by considering cross-decoder phone co-occurrences at the frame level, under two language modeling paradigms: smoothed n-grams and Support Vector Machines (SVM). Experiments on the NIST LRE2007 database demonstrated that using phone co-occurrence statistics could improve the perfor-mance of baseline phonotactic recognizers. In this paper, two variants of the cross-decoder phone co-occurrence SVM-based approach are proposed, by considering: (1) n-grams (up to 3-grams) of phone co-occurrences; and (2) co-occurrences of phone n-grams (up to 3-grams). To evaluate these approaches, a choice of open software (Brno University of Technology phone decoders, LIB-LINEAR and FoCal) was used, and experiments were carried out on the NIST LRE2007 database. Unlike those presented in [1], the two approaches presented in this pa-per outperformed the baseline phonotactic system, yield-ing around 16 % relative improvement in terms of EER. The best fused system attained a 1,88 % EER (a 30% improvement with regard to the baseline system), which supports the use of cross-decoder dependencies for lan-guage modeling. 1. |
| File Format | |
| Language | English |
| Publisher Date | 2010-01-01 |
| Access Restriction | Open |
| Subject Keyword | Open Software Cross-decoder Phone Co-occurrence Baseline System Svm-based Phonotactic Language Recognition Baseline Phonotactic Recognizers Frame Level Support Vector Machine Uncoupled Way Technology Phone Decoder Relative Improvement Fused System Nist Lre2007 Database Several Independent Phone Decod-ings Phone N-grams Phone Co-occurrence Ac-count Time Alignment Information New Approach Brno University Lan-guage Modeling Phonotactic Language Recognition Phone Co-occurrence Statistic Baseline Phonotactic System Time Alignment Common Approach Phonotactic Language Recognition Deal Cross-decoder Phone Co-occurrence Svm-based Approach Cross-decoder Dependency |
| Content Type | Text |
| Resource Type | Article |
