Academic Journal
Genones: generalized mixture tying in continuous hidden Markov model-based speech recognizers
العنوان: | Genones: generalized mixture tying in continuous hidden Markov model-based speech recognizers |
---|---|
المؤلفون: | Διγαλακης Βασιλης(http://users.isc.tuc.gr/~vdigalakis), Digalakis Vasilis(http://users.isc.tuc.gr/~vdigalakis), Monaco P.(), Murveit H.() |
بيانات النشر: | Institute of Electrical and Electronics Engineers |
سنة النشر: | 2015 |
المجموعة: | Technical University of Crete: Institutional Repository / Πολυτεχνείο Κρήτης: Ιδρυματικό Αποθετήριο |
مصطلحات موضوعية: | HMM,Hidden Markov Model, Speech recognition |
الوصف: | Summarization: An algorithm is proposed that achieves a good tradeoff between modeling resolution and robustness by using a new, general scheme for tying of mixture components in continuous mixture-density hidden Markov model (HMM)-based speech recognizers. The sets of HMM states that share the same mixture components are determined automatically using agglomerative clustering techniques. Experimental results on ARPA's Wall Street Journal corpus show that this scheme reduces errors by 25% over typical tied-mixture systems. New fast algorithms for computing Gaussian likelihoods-the-most time-consuming aspect of continuous-density HMM systems-are also presented. These new algorithms-significantly reduce the number of Gaussian densities that are evaluated with little or no impact on speech recognition accuracy ; Παρουσιάστηκε στο: IEEE Transactions Speech and Audio Processing |
نوع الوثيقة: | article in journal/newspaper |
اللغة: | English |
Relation: | info:eu-repo/grantAgreement/EC/FP7/246686; http://purl.tuc.gr/dl/dias/6BAF4883-036A-47B4-8A7B-A640F68EB6C9; http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=506931&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel4%2F89%2F11102%2F00506931 10.1109/89.506931 |
DOI: | 10.1109/89.506931 |
الاتاحة: | http://purl.tuc.gr/dl/dias/6BAF4883-036A-47B4-8A7B-A640F68EB6C9 http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=506931&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel4%2F89%2F11102%2F00506931 10.1109/89.506931 |
Rights: | info:eu-repo/semantics/openAccess ; Όροι χρήσης: Μη διαθέσιμο ; License: http://creativecommons.org/licenses/by/4.0/ |
رقم الانضمام: | edsbas.4EACA2BC |
قاعدة البيانات: | BASE |
DOI: | 10.1109/89.506931 |
---|