Search results for "speech recognition"
showing 10 items of 357 documents
Auditory Late Cortical Response and Speech Recognition in Digisonic Cochlear Implant Users
2002
Objective The purpose of the study was to test for differences in late electrically evoked auditory potentials between subjects exhibiting “good” versus “poor” speech recognition performances with their cochlear implants. Methods Late auditory evoked responses were measured in 30 subjects equipped with the Digisonic (MXM, Antibes, France) cochlear implant, 15 of whom had “good” speech recognition scores (i.e., more than 89% correct phoneme identification without lip reading). The 15 other subjects had poorer speech recognition scores (i.e., less than 85%). Results Differences in N1P2 amplitude, as well as P1, N1, and P2 latencies, and N1-P1 and N1-P2 latency intervals were tested. Wave P2 l…
Decoding Musical Training from Dynamic Processing of Musical Features in the Brain
2018
AbstractPattern recognition on neural activations from naturalistic music listening has been successful at predicting neural responses of listeners from musical features, and vice versa. Inter-subject differences in the decoding accuracies have arisen partly from musical training that has widely recognized structural and functional effects on the brain. We propose and evaluate a decoding approach aimed at predicting the musicianship class of an individual listener from dynamic neural processing of musical features. Whole brain functional magnetic resonance imaging (fMRI) data was acquired from musicians and nonmusicians during listening of three musical pieces from different genres. Six mus…
Key issues in decomposing fMRI during naturalistic and continuous music experience with independent component analysis
2014
Background: Independent component analysis (ICA) has been often used to decompose fMRI data mostly for the resting-state, block and event-related designs due to its outstanding advantage. For fMRI data during free-listening experiences, only a few exploratory studies applied ICA.New method: For processing the fMRI data elicited by 512-s modern tango, a FFT based band-pass filter was used to further pre-process the fMRI data to remove sources of no interest and noise. Then, a fast model order selection method was applied to estimate the number of sources. Next, both individual ICA and group ICA were performed. Subsequently, ICA components whose temporal courses were significantly correlated …
Surrogate data approaches to assess the significance of directed coherence: Application to EEG activity propagation
2009
This paper addresses the topic of evaluating the significance of frequency domain measures of causal coupling in multivariate time series through generation of surrogate data. The considered approaches are the traditional Fourier Transform (FT) algorithm and a new causal FT (CFT) algorithm for surrogate data generation. Both algorithms preserve the FT modulus of the original series; differences are in the phase relationships, that are completely destroyed for FT surrogates and imposed after switching off the link over the considered causal direction for CFT surrogates. The ability of the algorithms to assess causality in the frequency domain was tested using the directed coherence as discri…
Blocking by word frequency and neighborhood density in visual word recognition: A task-specific response criteria account
2004
International audience; Effects of blocking words by frequency class (high vs. low) and neighborhood density (high vs. low) were examined in two experiments using progressive demasking and lexical decision tasks. The aim was to examine the predictions of a task-specific response criteria account of list-blocking effects. Distinct patterns of blocking effects were obtained in the two tasks. In the progressive demasking task, a pure-list disadvantage was obtained to low frequency-high density words, whereas high frequency-low density produced a trend toward a pure-list advantage. In lexical decision, high-frequency words showed a pure-list advantage that was strongest in high-density words, w…
E-Hitz: A word frequency list and a program for deriving psycholinguistic statistics in an agglutinative language (Basque)
2007
We describe a Windows program that enables users to obtain a broad range of statistics concerning the properties of word and nonword stimuli in an agglutinative language (Basque), including measures of word frequency (at the whole-word and lemma levels), bigram and biphone frequency, orthographic similarity, orthographic and phonological structure, and syllable-based measures. It is designed for use by researchers in psycholinguistics, particularly those concerned with recognition of isolated words and morphology. In addition to providing standard orthographic and phonological neighborhood measures, the program can be used to obtain information about other forms of orthographic similarity, …
Sign Languages Recognition Based on Neural Network Architecture
2017
In the last years, many steps forward have been made in speech and natural languages recognition and were developed many virtual assistants such as Apple’s Siri, Google Now and Microsoft Cortana. Unfortunately, not everyone can use voice to communicate to other people and digital devices. Our system is a first step for extending the possibility of using virtual assistants to speech impaired people by providing an artificial sign languages recognition based on neural network architecture.
Room acoustical parameters: A factor analysis approach
2009
Abstract In this study, we determined the most representative acoustical parameters for halls intended for verbal or music audition. Our study was carried out in nine halls of different shapes and designed for different uses. We measured the impulse response at a great number of points (many more than the minimum required by the ISO 3382 norm). From a physical viewpoint, all halls are enclosed three-dimensional areas. Our work hypothesis is that objective (measurable) acoustical parameters, or a combination of such parameters, must provide the acoustical information specific to each hall and must make it possible to grade each hall. Factor analysis was used to obtain these grading parameter…
An offline/real-time artifact rejection strategy to improve the classification of multi-channel evoked potentials
2008
The primary goal of this paper is to improve the classification of multi-channel evoked potentials (EPs) by introducing a temporal domain artifact detection strategy and using this strategy to (a) evaluate how the performance of classifiers is affected by artifacts and (b) show how the performance can be improved by detecting and rejecting artifacts in offline and real-time classification experiments. Using a pattern recognition approach, an artifact is defined in this study as any signal that may lead to inaccurate classifier parameter estimation and inaccurate testing. The temporal domain artifact detection tests include: a within-channel standard deviation (STD) test that can detect sign…
A Study of Perceptron Mapping Capability to Design Speech Event Detectors
2006
Event detection is a fundamental yet critical component in automatic speech recognition (ASR) systems that attempt to extract knowledge-based features at the front-end level. In this context, it is common practice to design the detectors inside well-known frameworks based on artificial neural network (ANN) or support vector machine (SVM). In the case of ANN, speech scientists often design their detector architecture relying on conventional feed-forward multi-layer perceptron (MLP) with sigmoidal activation function. The aim of this paper is to introduce other ANN architectures inside the context of detection-based ASR. In particular, a bank of feed-forward MLPs using sinusoidal activation f…