ACOUSTIC-PHONETIC DECODING OF SPANISH CONTINUOUS SPEECH

1994 ◽  
Vol 08 (01) ◽  
pp. 155-180 ◽  
Author(s):  
I. GALIANO ◽  
E. SANCHIS ◽  
F. CASACUBERTA ◽  
I. TORRES
Keyword(s):  
Markov Models ◽  
Hidden Markov ◽  
Recognition Rate ◽  
Grammatical Inference ◽  
Mathematical Framework ◽  
Markov Modelling ◽  
Unit Model ◽  
Training Samples ◽  
Speech Corpora ◽  
Selection Of

The design of current acoustic-phonetic decoders for a specific language involves the selection of an adequate set of sublexical units, and a choice of the mathematical framework for modelling the corresponding units. In this work, the baseline chosen for continuous Spanish speech consists of 23 sublexical units that roughly correspond to the 24 Spanish phonemes. The process of selection of such a baseline was based on language phonetic criteria and some experiments with an available speech corpora. On the other hand, two types of models were chosen for this work, conventional Hidden Markov Models and Inferred Stochastic Regular Grammars. With these two choices we could compare classical Hidden Markov modelling where the structure of a unit-model is deductively supplied, with Grammatical Inference modelling where the baseforms of model-units are automatically generated from training samples. The best speaker-independent phone recognition rate was 64% for the first type of modelling, and 66% for the second type.

Recognition of Cursive Arabic Handwritten Text Using Embedded Training Based on Hidden Markov Models

2017 ◽  
Vol 32 (01) ◽  
pp. 1860007 ◽  
Author(s):  
Mouhcine Rabi ◽  
Mustapha Amrouch ◽  
Zouhir Mahani
Keyword(s):  
Hidden Markov Models ◽  
Markov Models ◽  
Hidden Markov ◽  
Recognition Rate ◽  
Distribution Characteristics ◽  
Handwritten Text ◽  
Baseline Estimation

This paper presents a system for offline recognition of cursive Arabic handwritten text based on Hidden Markov Models (HMMs). The proposed work reports an effective method taking into account the context of character by applying an embedded training-based HMMs to perform and enhance the character models. The system is analytical without explicit segmentation; extracted features preceded by baseline estimation are statistical and structural to integrate both the peculiarities of the text and the pixel distribution characteristics of the word image. The experiments are done on benchmark IFN/ENIT database. The proposed work shows the effectiveness of using embedded training-based HMMs for enhancing the recognition rate, and the obtained results are promising and encouraging.

A Comparative Study on Feature Selection of Text Categorization for Hidden Markov Models

2013 ◽  
Author(s):  
Kwan Yi ◽  
Jamshid Beheshti
Keyword(s):  
Feature Selection ◽  
Text Categorization ◽  
Markov Models ◽  
Hidden Markov ◽  
Model Performance ◽  
Document Representation ◽  
Selection Methods ◽  
Learning Models ◽  
Text Feature ◽  
Selection Of

In document representation for digitalized text, feature selection refers to the selection of the terms of representing a document and of distinguishing it from other documents. This study probes different feature selection methods for HMM learning models to explore how they affect the model performance, which is experimented in the context of text categorization task.Dans la représentation documentaire des textes numérisés, la sélection des caractéristiques se fonde sur la sélection des termes représentant et distinguant un document des autres documents. Cette étude examine différents modèles de sélection de caractéristiques pour les modèles d’apprentissage MMC, afin d’explorer comment ils affectent la performance du modèle, qui est observé dans le contexte de la tâche de catégorisation textuelle. 

CONSTRAINED LEARNING VECTOR QUANTIZATION

1994 ◽  
Vol 05 (02) ◽  
pp. 143-152
Author(s):  
HONG YAN
Keyword(s):  
Vector Quantization ◽  
Nearest Neighbor ◽  
Recognition Performance ◽  
Recognition Rate ◽  
Learning Vector Quantization ◽  
Training Time ◽  
Training Samples ◽  
Constrained Learning ◽  
Neighbor Classifier ◽  
Selection Of

Kohonen’s learning vector quantization (LVQ) is an efficient neural network based technique for pattern recognition. The performance of the method depends on proper selection of the learning parameters. Over-training may cause a degradation in recognition rate of the final classifier. In this paper we introduce constrained learning vector quantization (CLVQ). In this method the updated coefficients in each iteration are accepted only if the recognition performance of the classifier after updating is not decreased for the training samples compared with that before updating, a constraint widely used in many prototype editing procedures to simplify and optimize a nearest neighbor classifier (NNC). An efficient computer algorithm is developed to implement this constraint. The method is verified with experimental results. It is shown that CLVQ outperforms and may even require much less training time than LVQ.

scholarly journals Fitting occupancy models with E-SURGE: Hidden Markov modelling of presence-absence data

2013 ◽  
Author(s):  
Olivier Gimenez ◽  
Laetitia Blanc ◽  
Aurélien Besnard ◽  
Roger Pradel ◽  
Paul Doherty ◽  
...  
Keyword(s):  
Markov Models ◽  
Hidden Markov ◽  
Numerical Algorithms ◽  
Simulated Data ◽  
Model Specification ◽  
Occupancy Models ◽  
Markov Modelling ◽  
State Process ◽  
Observation Process ◽  
Capture Recapture

1. Occupancy – the proportion of area occupied by a species – is a key notion for addressing important questions in ecology, biogeography and conservation biology. Occupancy models allow estimating and inferring about species occurrence while accounting for false absences (or imperfect species detection). 2. Most occupancy models can be formulated as hidden Markov models (HMM) in which the state process captures the Markovian dynamic of the actual but latent states while the observation process consists of observations that are made from these underlying states. 3. We show how occupancy models can be implemented in program E-SURGE, which was initially developed to analyse capture-recapture data in the HMM framework. Replacing individuals by sites provides the user with access to several features of E-SURGE that are not available altogether or just not available in standard occupancy software: i) user-friendly model specification through a SAS/R-like syntax without having to write custom code, ii) decomposition of the observation and state processes in several steps to provide flexible parameterisation, iii) up-to-date diagnostics of model identifiability and iv) advanced numerical algorithms to produce fast and reliable results (including site random effects). 4. To illustrate E-SURGE features, we provide simulated data and the details of the implementation on the analysis of several occupancy models. These detailed examples are gathered in a companion wiki platform http://occupancyinesurge.wikidot.com/ .

scholarly journals Application of Random Forest and Hidden Markov Models for Automatic and Fast Classification of Power Quality Signals

2020 ◽  
Vol 9 (4) ◽  
pp. 2856-2862
Keyword(s):  
Wavelet Transform ◽  
Random Forest ◽  
Wavelet Transforms ◽  
Markov Models ◽  
Hidden Markov ◽  
Recognition Rate ◽  
Discrete Wavelet ◽  
Second Generation Wavelet Transform ◽  
Quality Signals

In this paper, wavelet transform, namely the maximal overlap discrete Wavelet Transform (MODWT) and the second generation Wavelet Transform (SGWT) have been implemented. These wavelet transforms are applied to get selected features of the signals. Features are used as inputs to two types of classifiers namely, Hidden Markov Model (HMM) classifiers and the Random Forest (RF) classifier in the both absence and presence of Noise to evaluate the efficiency. The classification accuracy (CA) calculated using these classifiers clearly shows that the RF classifiers is a better classifier then the HMM classifier as it possess higher recognition rate at all levels of noise along with the pure PQ signals. Another important property of RF classifier is the proper classification of large number of class of both slow and the fast disturbances.

scholarly journals LAPPED TRANSFORMS AND HIDDEN MARKOV MODELS FOR SEISMIC DATA FILTERING

2004 ◽  
Vol 02 (04) ◽  
pp. 455-476
Author(s):  
LAURENT DUVAL ◽  
CAROLINE CHAUX
Keyword(s):  
Hidden Markov Models ◽  
Seismic Data ◽  
Markov Models ◽  
Hidden Markov ◽  
Image Features ◽  
Seismic Exploration ◽  
Pass Band ◽  
Markov Modelling ◽  
Noise Sources ◽  
Block Transforms

Seismic exploration provides information about the ground substructures. Seismic images are generally corrupted by several noise sources. Hence, efficient denoising procedures are required to improve the detection of essential geological information. Wavelet bases provide sparse representation for a wide class of signals and images. This property makes them good candidates for efficient filtering tools, allowing the separation of signal and noise coefficients. Recent works have improved their performance by modelling the intra- and inter-scale coefficient dependencies using hidden Markov models, since image features tend to cluster and persist in the wavelet domain. This work focuses on the use of lapped transforms associated with hidden Markov modelling. Lapped transforms are traditionally viewed as block-transforms, composed of M pass-band filters. Seismic data present oscillatory patterns and lapped transforms oscillatory bases have demonstrated good performances for seismic data compression. A dyadic like representation of lapped transform coefficient is possible, allowing a wavelet-like modelling of coefficients dependencies. We show that the proposed filtering algorithm often outperforms the wavelet performance both objectively (in terms of SNR) and subjectively: lapped transform better preserve the oscillatory features present in seismic data at low to moderate noise levels.

Fuzzy Observable Markov Models for Pattern Recognition

2007 ◽  
Vol 11 (6) ◽  
pp. 662-667
Author(s):  
Dat Tran ◽  
◽  
Wanli Ma ◽  
Dharmendra Sharma
Keyword(s):  
Pattern Recognition ◽  
Markov Model ◽  
Hidden Markov Model ◽  
Vector Quantization ◽  
Markov Models ◽  
Hidden Markov ◽  
Gaussian Mixture ◽  
Written Language ◽  
Mathematical Framework ◽  
Fuzzy Observable

This paper presents a mathematical framework for fuzzy discrete and continuous observable Markov models (OMMs) and their applications to written language, spam email and typist recognition. Experimental results show that the proposed OMMs are more effective than models such as vector quantization, Gaussian mixture model and hidden Markov model.

scholarly journals Hidden-Markov program algebra with iteration

2014 ◽  
Vol 25 (2) ◽  
pp. 320-360 ◽  
Author(s):  
ANNABELLE MCIVER ◽  
LARISSA MEINICKE ◽  
CARROLL MORGAN
Keyword(s):  
Markov Models ◽  
Hidden Markov ◽  
Information Leakage ◽  
Program Algebra ◽  
Password Guessing Attack ◽  
Implementation Program ◽  
Algebraic Laws ◽  
Source Level ◽  
Guessing Attack ◽  
Selection Of

We use hidden Markov models to motivate a quantitative compositional semantics for noninterference-based security with iteration, including a refinement- or ‘implements’ relation that compares two programs with respect to their information leakage; and we propose a program algebra for source-level reasoning about such programs, in particular as a means of establishing that an ‘implementation’ program leaks no more than its ‘specification’ program.This joins two themes: we extend our earlier work, having iteration but only qualitative (Morgan 2009), by making it quantitative; and we extend our earlier quantitative work (McIver et al. 2010) by including iteration.We advocate stepwise refinement and source-level program algebra – both as conceptual reasoning tools and as targets for automated assistance. A selection of algebraic laws is given to support this view in the case of quantitative noninterference; and it is demonstrated on a simple iterated password-guessing attack.

scholarly journals Multimodal face recognition method with two-dimensional hidden Markov model

2017 ◽  
Vol 65 (1) ◽  
pp. 121-128 ◽  
Author(s):  
J. Bobulski
Keyword(s):  
Face Recognition ◽  
Hidden Markov Models ◽  
Markov Models ◽  
Hidden Markov ◽  
Recognition Rate ◽  
Two Dimensional ◽  
One Dimensional ◽  
3D Face ◽  
Multimodal Face Recognition ◽  
2D And 3D

Abstract The paper presents a new solution for the face recognition based on two-dimensional hidden Markov models. The traditional HMM uses one-dimensional data vectors, which is a drawback in the case of 2D and 3D image processing, because part of the information is lost during the conversion to one-dimensional features vector. The paper presents a concept of the full ergodic 2DHMM, which can be used in 2D and 3D face recognition. The experimental results demonstrate that the system based on two dimensional hidden Markov models is able to achieve a good recognition rate for 2D, 3D and multimodal (2D+3D) face images recognition, and is faster than ICP method.

Sign in / Sign up

Export Citation Format

Share Document