Speaker Adaptation Applied to Sinhala Speech Recognition
نویسندگان
چکیده
Sinhala, which the main spoken language of the majority of Sri Lanka, is an under-resourced language. Sinhala language is new to the speech recognition research field and faces the problem of not having suitable speech corpora available. For a language like Sinhala, it is essential to find out ways of developing good recognition models using a fewer sample of data. Speaker Adaptive methods provides the opportunity of improving speaker independent recognition systems into more speaker dependent systems by adapting the features of the user. In this paper we are presenting an experiment we carried out by adapting a pre-trained Sinhala speech recognition system (with a single voice) with several different speaker voices. Our experiment shows that although individual adaptation systems gives the best results for corresponding speakers, we can build general speaker adaptation models to get better results than building speaker independent models using the same amount of data.
منابع مشابه
Speaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملSpeaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملRobust Speech Recognition Usin Intra-speaker Ada
Inter-speaker variation can be coped rather well in speech recognition by speaker adaptation techniques such as MLLR and MAP. However, when dealing with speech other than reading style, such as conversational speech, emotional speech and so on, current recognition systems cannot achieve a satisfactory performance even after speaker adaptation. In view of this situation, two-level adaptation met...
متن کاملA Comparative Study of Gender and Age Classification in Speech Signals
Accurate gender classification is useful in speech and speaker recognition as well as speech emotion classification, because a better performance has been reported when separate acoustic models are employed for males and females. Gender classification is also apparent in face recognition, video summarization, human-robot interaction, etc. Although gender classification is rather mature in a...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- Int. J. Comput. Linguistics Appl.
دوره 6 شماره
صفحات -
تاریخ انتشار 2015