Automatic Prosody Labeling Final Project Report for EE 6820 - Spring 05 Professor : Dan
نویسندگان
چکیده
Automatic transcription of prosody is necessary for spoken language understanding. Prominence and intonational boundaries are routinely used to convey meaning beyond that expressed in the lexical content of speech. Using a classiÞcation rule learning algorithm and computationally light acoustic and syntactic features, detection of pitch accent at 87% on spontaneous elicited speech were attained along with 94% accurate detection of full intonational phrase boundaries.
منابع مشابه
Identification E 6820 Spring ’ 08 Final Project Report Prof . Dan Ellis
People use biometric information to distinguish between different persons. Visually, face is one most important feature, other unique features, such as finger-prints, iris, are often used. Another way to identify a person is from the acoustic fact that each person’s voice are different, this forms one area of speech processing, automatic speaker recognition. For the past few decades, many solut...
متن کاملThe University of Washington , Department of EE Technical Report Series
Automatic annotation of prosodic events could help improve speech understanding and synthesis. However, little annotated data is available for training prosody models because hand-labeling is prohibitively expensive. To address this issue, we explore weakly supervised learning techniques (EM, co-training, and self-training with bagging) that use only a small amount of hand-labeled data in combi...
متن کاملPerceptually-Related F0 Parameters for Automatic Classification of Phrase Final Tones
Automatic labeling of prosodic features is an important topic when constructing large speech databases for speech synthesis or analysis purposes. Perceptually-related F0 parameters are proposed with the aim of automatically classifying phrase final tones. Analyses are conducted to verify how consistently subjects are able to categorize phrase final tones, and how perceptual features are related...
متن کاملAutomatic labeling of prosody
The paper proposes a framework for automatic prosody labeling. The labeling involves detection of the location of accented syllables and phrase boundaries, and recognition of pitch accent and boundary tone types. A number of classification models are designed to perform these tasks on the basis of small vectors of acoustic features. The models achieve high accuracy and their performance is comp...
متن کاملAutomatic labeling of Japanese prosody using j-toBI style description
Speech corpora with prosodic labels are getting more and more important not only for speech synthesis but also for discourse modeling. A widely used labeling system for Japanese prosody, J-ToBI, however, is insufficient for applications like discourse modeling and it even lacks an accurate method for automatic labeling. In this paper, we propose an automatic labeling method for J-ToBI style des...
متن کامل