Understanding speech in noisy environments is challenging, especially for seniors. Although evidence suggests that older adults increasingly recruit prefrontal cortices to offset reduced periphery and central auditory processing, the brain mechanisms underlying such compensation remain elusive. Here we show that relative to young adults, older adults show higher activation of frontal speech mot...

We build an automatic phoneme recognition system based on Bayesian Multivariate Modeling which is a static scheme. Phoneme models were built by using stochastic pattern recognition and acoustic phonetic schemes to recognise phonemes. Since our native language is Kannada, a rich South Indian Language, we have used 15 Kannada phonemes to train and test these models. As Mel – Frequency Cepstral Co...

Research into human perception of consonants has identified phoneme-specific perceptual cues. It has also been shown that the characteristics of the speech signal most useful for recognition depend on the specific speech sound. Typical ASR features and recognisers however neither vary with the type of sound nor relate directly to perceptual cues. We investigate classification and decoding of no...

The contribution addresses the cross-language experiment. The aim was to test the possibility of the conversion French phoneme models into Czech ones. This model conversion uses the Hidden Markov Models (HMM) classification procedure. The first step consists of the iterative mapping of French models to Czech ones. The mapping is given by the analysis the confusion matrix. The second step is the...

In this work, new multi-classifier schemes for isolated word speech recognition based on the combination of standard Hidden Markov Models (HMMs) and Complementary Gaussian Mixture Models (CGMMs) are proposed. Typically, in speech recognition systems, each word or phoneme in the vocabulary is represented by a model trained with samples of each particular class. The recognition is then performed ...

It is believed that knowledge gained from reliable accent classification could be employed to improve the performance of speech recognition and speaker recognition algorithms. In this paper, we investigate the use of articulatory movement in the spectral domain to classify accented speech. The trajectories are modelled by a mixture of probability density functions of a random sequence of states...

We introduce both neural higher-order linear-chain conditional random fields (NHO-LC-CRFs) and a new structured regularizer for these sequence models. We show that this regularizer can be derived as lower bound from a mixture of models sharing parts of each other, e.g. neural sub-networks, and relate it to ensemble learning. Furthermore, it can be expressed explicitly as regularization term in ...

This paper proposes a new electromyogram (EMG) pattern classification method using probabilistic neuralnetworks based on boosting approach [1]. Since the proposed method automatically constructs a suitable classification network from measured EMG signals, there is no need to set the structure of network in advance. To verify the feasibility of the proposed method, phoneme classification experim...