Enhanced discriminant analysis for confusable sounds via speaker adaptation techniques
نویسنده
چکیده
Speaker adaptation is important to reduce the variability of the speech signal from various speakers. Two approaches are introduced to solve this problem, Vector Quantization (VQ) and Canonical Correlation (CC). Their ability to enhance the discrimination of confused sounds are evaluated through a collected database of Arabic confusable sounds. A recognition accuracy of 73% is obtained without any speaker adaptation technique. When the VQ is applied the accuracy increases to 88%. When the CC Technique is applied the accuracy increases to 93%. This proves the ability of the Canonical Correlation analysis versus VQ.
منابع مشابه
Gender Clustering Improvement for Speaker Adaptation Using LDA
Speaker adaptation is an important issue in current speech recognition researches. Speaker clustering is one of the widely used methods in speaker adaptation. In this paper the effect of Linear Discriminant Analysis (LDA) on increasing the accuracy of some clustering methods such as k-means and Support Vector Machine (SVM) is demonstrated. The performance of this idea was examined on AURORA 2.0...
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 کاملExtended Variability Modeling and Unsupervised Adaptation for PLDA Speaker Recognition
Probabilistic Linear Discriminant Analysis (PLDA) continues to be the most effective approach for speaker recognition in the i-vector space. This paper extends the PLDA model to include both enrollment and test cut duration as well as to distinguish between session and channel variability. In addition, we address the task of unsupervised adaptation to unknown new domains in two ways: speaker-de...
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 کاملEliminating Inter-speaker Variability Prior to Discriminant Transforms
This paper shows the impact of speaker normalization techniques such as vocal tract length normalization (VTLN) and speaker-adaptive training (SAT) prior to discriminant feature space transforms, such as LDA. We demonstrate that removing the inter-speaker variability by using speaker compensation methods results in improved discrimination as measured by the LDA eigenvalues and also in improved ...
متن کامل