Continuous Density Hidden Markov Models Research Articles

651 Longitudinal, unobtrusive, and ecologically valid sleep metric estimation from a smart bed to predict the pathology of COVID-19

IntroductionPathophysiologic responses to viral respiratory challenges such as SARS-CoV-2 may affect sleep duration, quality and concomitant cardiorespiratory function. Unobtrusive and ecologically valid methods to monitor longitudinal sleep metrics may therefore have practical value for surveillance and monitoring of infectious illnesses. We leveraged sleep metrics from Sleep Number 360 smart bed users to build a COVID-19 predictive model.MethodsAn IRB approved survey was presented to opting-in users from August to November 2020. COVID-19 test results were reported by 2003/6878 respondents (116 positive; 1887 negative). From the positive group, data from 82 responders (44.7±11.3 yrs.) who reported the date of symptom onset were used. From the negative group, data from 1519 responders (48.4±12.9 yrs.) who reported testing dates were used. Sleep duration, sleep quality, restful sleep duration, time to fall asleep, respiration rate, heart rate, and motion level were obtained from ballistocardiography signals stored in the cloud. Data from January to October 2020 were considered. The predictive model consists of two levels: 1) the daily probability of staying healthy calculated by logistic regression and 2) a continuous density Hidden Markov Model to refine the daily prediction considering the past decision history.ResultsWith respect to their baseline, significant increases in sleep duration, average breathing rate, average heart rate and decrease in sleep quality were associated with symptom exacerbation in COVID-19 positive respondents. In COVID-19 negative respondents, no significant sleep or cardiorespiratory metrics were observed. Evaluation of the predictive model resulted in cross-validated area under the receiving-operator curve (AUC) estimate of 0.84±0.09 which is similar to values reported for wearable-sensors. Considering additional days to confirm prediction improved the AUC estimate to 0.93±0.05.ConclusionThe results obtained on the smart bed user population suggest that unobtrusive sleep metrics may offer rich information to predict and track the development of symptoms in individuals infected with COVID-19.Support (if any):

A Novel Research in Low Altitude Acoustic Target Recognition Based on HMM

This paper introduces an improved HMM (hidden Markov model) for low altitude acoustic target recognition. To overcome the limitation of the classical CDHMM (continuous density hidden Markov model) training algorithm and the generalization ability deficiency of existing discriminative learning methods, a new discriminative training method for estimating the CDHMM in acoustic target recognition is proposed based on the principle of maximizing the minimum relative separation margin. According to the definition of the relative margin, the new training criterion can be equation as a standard constrained minimax optimization problem. Then, the optimization problem can be solved by a GPD (generalized probabilistic descent) algorithm. The experimental results show that the performance of the algorithm is significantly improved compared with the former training method, which can effectively improve the recognition ability of the acoustic target recognition system.

Hybrid Continuous Density Hmm-Based Ensemble Neural Networks for Sensor Fault Detection and Classification in Wireless Sensor Network.

Sensor devices in wireless sensor networks are vulnerable to faults during their operation in unmonitored and hazardous environments. Though various methods have been proposed by researchers to detect sensor faults, only very few research studies have reported on capturing the dynamics of the inherent states in sensor data during fault occurrence. The continuous density hidden Markov model (CDHMM) is proposed in this research to determine the dynamics of the state transitions due to fault occurrence, while neural networks are utilized to classify the faults based on the state transition probability density generated by the CDHMM. Therefore, this paper focuses on the fault detection and classification using the hybridization of CDHMM and various neural networks (NNs), namely the learning vector quantization, probabilistic neural network, adaptive probabilistic neural network, and radial basis function. The hybrid models of each NN are used for the classification of sensor faults, namely bias, drift, random, and spike. The proposed methods are evaluated using four performance metrics which includes detection accuracy, false positive rate, F1-score, and the Matthews correlation coefficient. The simulation results show that the learning vector quantization NN classifier outperforms the detection accuracy rate when compared to the other classifiers. In addition, an ensemble NN framework based on the hybrid CDHMM classifier is built with majority voting scheme for decision making and classification. The results of the hybrid CDHMM ensemble classifiers clearly indicates the efficacy of the proposed scheme in capturing the dynamics of change of statesm which is the vital aspect in determining rapidly-evolving instant faults that occur in wireless sensor networks.

N-Gram Language Model based Continuous Voiced Odia Digit Recognition

With the enormous improvement in the area of signal processing, speech processing systems are creating a massive impact in recognizing the voices, controlling the commands and making as communication interfaces. A continuous speech recognition system is essential for voice identification hands free system used as a voice dialer, voice originated security systems and voice based automatic electronic machines. The proposed work suggests a finest speaker independent continuous voiced digit recognition for Odia language. The model integrates the concept of Mel Frequency Cepstral Coefficient (MFCC) and continuous density Hidden Markov Model (HMM), relating to speech parameterization and recognition respectively. The performance of the model is explored for different levels of HMM like word-level and phoneme-level. Further the model output is evaluated using different N-Gram approaches of the language model. Finally it is shown that the model using phoneme-level HMM with a tri-gram language model is superior to other methodologies.

A Markov-Switching Model Approach to Heart Sound Segmentation and Classification

We consider challenges in accurate segmentation of heart sound signals recorded under noisy clinical environments for subsequent classification of pathological events. Existing state-of-the-art solutions to heart sound segmentation use probabilistic models such as hidden Markov models (HMMs), which, however, are limited by its observation independence assumption and rely on pre-extraction of noise-robust features. We propose a Markov-switching autoregressive (MSAR) process to model the raw heart sound signals directly, which allows efficient segmentation of the cyclical heart sound states according to the distinct dependence structure in each state. To enhance robustness, we extend the MSAR model to a switching linear dynamic system (SLDS) that jointly model both the switching AR dynamics of underlying heart sound signals and the noise effects. We introduce a novel algorithm via fusion of switching Kalman filter and the duration-dependent Viterbi algorithm, which incorporates the duration of heart sound states to improve state decoding. Evaluated on Physionet/CinC Challenge 2016 dataset, the proposed MSAR-SLDS approach significantly outperforms the hidden semi-Markov model (HSMM) in heart sound segmentation based on raw signals and comparable to a feature-based HSMM. The segmented labels were then used to train Gaussian-mixture HMM classifier for identification of abnormal beats, achieving high average precision of 86.1% on the same dataset including very noisy recordings. The proposed approach shows noticeable performance in heart sound segmentation and classification on a large noisy dataset. It is potentially useful in developing automated heart monitoring systems for pre-screening of heart pathologies.

Cochlea-inspired speech recognition interface.

Automatic speech recognition (ASR) technology provides a natural interface for human-machine interaction. Typical ASR systems can achieve high performance in quiet environments but, unlike humans, perform poorly in real-world situations. To better simulate the human auditory periphery and improve the performance in realistic noisy scenarios, we propose two models of speech recognition front-ends based on a biophysical cochlear model. The first front-end is based on the method of signal reconstruction from a basilar membrane response. When applied to noisy speech, this method results in improved signal quality. This method can be used as a preprocessing step in a standard ASR system and can also be used as a noise reduction technique for other applications. The second front-end we propose is based on the construction of speech recognition coefficients directly from a basilar membrane response. Experimental results using a continuous-density hidden Markov model (HMM) recognizer demonstrate significant improvement in performance compared to standard Mel-frequency cepstral coefficients (MFCC) in various types of noisy conditions. Graphical Abstract Speech recognition model based on cochlear front-end.

Robust Cochlear-Model-Based Speech Recognition

Accurate speech recognition can provide a natural interface for human–computer interaction. Recognition rates of the modern speech recognition systems are highly dependent on background noise levels and a choice of acoustic feature extraction method can have a significant impact on system performance. This paper presents a robust speech recognition system based on a front-end motivated by human cochlear processing of audio signals. In the proposed front-end, cochlear behavior is first emulated by the filtering operations of the gammatone filterbank and subsequently by the Inner Hair cell (IHC) processing stage. Experimental results using a continuous density Hidden Markov Model (HMM) recognizer with the proposed Gammatone Hair Cell (GHC) coefficients are lower for clean speech conditions, but demonstrate significant improvement in performance in noisy conditions compared to standard Mel-Frequency Cepstral Coefficients (MFCC) baseline.

Robust emotion recognition from speech: Gamma tone features and models

Affective computing is gaining paramount importance in ensuring the better and effective human–machine interaction. As glottal and speech signals depict the characteristics of the emotional nature of the speaker in addition to the linguistic information, speaker’s emotions are needed to be recognised to give meaningful response by the system. This paper emphasises the effectiveness and efficiency in selecting the energy features by passing the speech through the Gamma tone filters spaced in Equivalent rectangular bandwidth (ERB), MEL and BARK scale. Various modelling techniques are used to develop the robust multi-speaker independent speaker’s emotion/stress recognition system. Since EMO-DB Berlin database and SAVEE emotional audio-visual database used in this work contain the only limited set of speech utterances uttered by 10/4 actors/speakers in different emotions, it has become challenging to improve the performance of the stress/emotion recognition system. Speaker independent emotion recognition is done by extracting the Gamma tone energy features and cepstral features by passing the concatenated speech considered for training through the Gamma tone filters spaced in ERB, MEL and BARK scales. Subsequently, VQ/Fuzzy clustering models and continuous density hidden Markov models are created for all emotions and evaluation is done with the utterances of a speaker independent of speeches considered for training. The proposed features for test utterances are captured and applied to the VQ/Fuzzy/MHMM/SVM models and testing is performed by using minimum distance criterion/maximum log-likelihood criterion. The proposed Gamma tone energy/cepstral features and modelling techniques provide complementary evidence in assessing the performance of the system. This algorithm offers 96%, 79%, and 95.3% as weighted accuracy recall for the stress recognition system with respect to the classification done on emotion-specific group VQ/Fuzzy/MHMM/SVM models for GTF energy features with Gamma tone filters spaced in ERB, MEL and BARK scale respectively for the system evaluated for the EMO-DB database. Weighted accuracy recall is found to be 91%, 93% and 94% for the classification done on emotion-specific group models for GTF energy features with Gamma tone filters spaced in ERB, MEL and BARK scale respectively for the evaluation done on the utterances chosen from the SAVEE database. Gamma tone Cepstral features provide the overall accuracy of 92%, 90% and 92% for filters spaced in ERB, MEL and BARK scale for Berlin EMO-DB. Decision level fusion classification based on GTF energy features and modelling techniques provides the overall accuracy as 99.8% for EO-DB database and 100% for SAVEE database.

QRS complex detection based on continuous density hidden Markov models using univariate observations

In the electrocardiogram (ECG), the detection of QRS complexes is a fundamental step in the ECG signal processing chain since it allows the determination of other characteristics waves of the ECG and provides information about heart rate variability. In this work, an automatic QRS complex detector based on continuous density hidden Markov models (HMM) is proposed. HMM were trained using univariate observation sequences taken either from QRS complexes or their derivatives. The detection approach is based on the log-likelihood comparison of the observation sequence with a fixed threshold. A sliding window was used to obtain the observation sequence to be evaluated by the model. The threshold was optimized by receiver operating characteristic curves. Sensitivity (Sen), specificity (Spc) and F1 score were used to evaluate the detection performance. The approach was validated using ECG recordings from the MIT-BIH Arrhythmia database. A 6-fold cross-validation shows that the best detection performance was achieved with 2 states HMM trained with QRS complexes sequences (Sen = 0.668, Spc = 0.360 and F1 = 0.309). We concluded that these univariate sequences provide enough information to characterize the QRS complex dynamics from HMM. Future works are directed to the use of multivariate observations to increase the detection performance.

Sparse coding based features for speech units classification

In this work, we propose sparse representation based features for speech units classification tasks. In order to effectively capture the variations in a speech unit, the proposed method employs multiple class specific dictionaries. Here, the training data belonging to each class is clustered into multiple clusters, and a principal component analysis (PCA) based dictionary is learnt for each cluster. It has been observed that coefficients corresponding to middle principal components can effectively discriminate among different speech units. Exploiting this observation, we propose to use a transformation function known as weighted decomposition (WD) of principal components, which is used to emphasize the discriminative information present in the PCA-based dictionary. In this paper, both raw speech samples and mel frequency cepstral coefficients (MFCC) are used as an initial representation for feature extraction. For comparison, various popular dictionary learning techniques such as K-singular value decomposition (KSVD), simultaneous codeword optimization (SimCO) and greedy adaptive dictionary (GAD) are also employed in the proposed framework. The effectiveness of the proposed features is demonstrated using continuous density hidden Markov model (CDHMM) based classifiers for (i) classification of isolated utterances of E-set of English alphabet, (ii) classification of consonant-vowel (CV) segments in Hindi language and (iii) classification of phoneme from TIMIT phonetic corpus.

TAMEEM V1.0: speakers and text independent Arabic automatic continuous speech recognizer

This research work aims to disseminate the efforts towards developing and evaluating TAMEEM V1.0, which is a state-of-the-art pure Modern Standard Arabic (MSA), automatic, continuous, speaker independent, and text independent speech recognizer using high proportion of the spoken data of the phonetically rich and balanced MSA speech corpus. The speech corpus contains speech recordings of Arabic native speakers from 11 Arab countries representing Levant, Gulf, and Africa regions of the Arabic World, which make about 45.30 h of speech data. The recordings contain about 39.28 h of 367 sentences that are considered phonetically rich and balanced, which are used for training TAMEEM V1.0 speech recognizer, and another 6.02 h of another 48 sentences that are used for testing purposes, which are mostly text independent and foreign to the training sentences. TAMEEM V1.0 speech recognizer is developed using the Carnegie Mellon University (CMU) Sphinx 3 tools in order to evaluate the speech corpus, whereby the speech engine uses three-emitting state Continuous Density Hidden Markov Model for tri-phone based acoustic models, and the language model contains uni-grams, bi-grams, and tri-grams. Using three different testing data sets, this work obtained 7.64% average Word Error Rate (WER) for speakers dependent with text independent data set. For speakers independent with text dependent data set, this work obtained 2.22% average WER, whereas 7.82% average WER is achieved for speakers independent with text independent data set.

ECG segmentation and fiducial point extraction using multi hidden Markov model

In this paper, we propose a novel method for extracting fiducial points (FPs) of electrocardiogram (ECG) signals. We propose the use of multi hidden Markov model (MultiHMM) as opposed to the traditional use of Classic HMM. In the MultiHMM method, each segment of an ECG beat is represented by a separate ergodic continuous density HMM. Each HMM has different state number and is trained separately. In the test step, the log-likelihood of two consecutive HMMs is compared and a path is estimated, which shows the correspondence of each part of the ECG signal to the HMM with the maximum log-likelihood. Fiducial points are estimated from the obtained path. For performance evaluation, the Physionet QT database and a Swine ECG database are used and the proposed method is compared with the Classic HMM and a method based on partially collapsed Gibbs sampler (PCGS). In our evaluation using the QT database, we also compare the results with low-pass differentiation, hybrid feature extraction algorithm, a method based on the wavelet transform and three HMM-based approaches. For the Swine database, the root mean square error (RMSE) values, across all FPs for MultiHMM, Classic HMM and PCGS methods are 13, 21 and 40ms, respectively and the MultiHMM exhibits smaller error variability than other methods. For the QT database, RMSE values for MultiHMM, Classic HMM, Wavelet and PCGS methods are 10, 17, 26 and 38ms, respectively. Our results demonstrate that our proposed MultiHMM approach outperforms other benchmark methods that exist in the literature; therefore can be used in practical ECG fiducial point extraction.

Integrated models and features-based speaker independent emotion recognition

Speech emotion recognition has become a challenging task in speech technology in order to ensure better and effective human-machine interaction. Improving the accuracy of speaker independent emotion recognition system by using a database containing the speech samples of limited number of utterances spoken by limited set of speakers is more challenging. This paper mainly discusses the use of integrated features and integrated models for improving the accuracy of the emotion recognition system. Integrated features are obtained by concatenating the probability with the normal system features. Integrated models are obtained by combining clustering and continuous density hidden Markov models. Integrated models with Mel frequency perceptual linear predictive cepstrum concatenated with probability provides better accuracy of 89% for recognising emotions using EMO-DB which contains speech samples of ten different utterances of ten different speakers in different emotions such as anger, boredom, disgust, fear, happy, neutral and sad by applying spectral analysis as a additional pre processing technique.

Unsupervised Learning of Continuous Density HMM for Variable-Length Spoken Unit Discovery

Unsupervised Learning of Continuous Density HMM for Variable-Length Spoken Unit Discovery

Full body movements recognition – unsupervised learning approach with heuristic R-GDL method

In computer systems that are used for actions recognition the human movements are often represented by three-dimensional coordinates of body joints that are tracked by motion capture hardware. The motivation of our research was to propose a novel method for automatic generation of knowledge base for syntactic Gesture Description Language (GDL) classifier by analyzing unsegmented data recordings of gestures. We have proposed novel unsupervised learning approach to deal with this task. Because this process seems to be reverse engineering to GDL approach, the learning algorithm we introduce in this paper, is called Revers-GDL (R-GDL). The R-GDL machine-learning approach for full-body movements recognition is a novel method of time-varying multidimensional signals classification. The description of R-GDL and its validation is our original and never before published achievement. The evaluation of R-GDL was performed with k-fold cross validation on large dataset that contains 770 complete movements samples of 9 gym exercises performed by 14 persons and compared with results from multivariate normally continuous density hidden Markov model classifier. Depending on exercise type GDL obtained recognition rate at the level of 100% to 91%.

A Recurrent Probabilistic Neural Network with Dimensionality Reduction Based on Time-series Discriminant Component Analysis.

This paper proposes a probabilistic neural network (NN) developed on the basis of time-series discriminant component analysis (TSDCA) that can be used to classify high-dimensional time-series patterns. TSDCA involves the compression of high-dimensional time series into a lower dimensional space using a set of orthogonal transformations and the calculation of posterior probabilities based on a continuous-density hidden Markov model with a Gaussian mixture model expressed in the reduced-dimensional space. The analysis can be incorporated into an NN, which is named a time-series discriminant component network (TSDCN), so that parameters of dimensionality reduction and classification can be obtained simultaneously as network coefficients according to a backpropagation through time-based learning algorithm with the Lagrange multiplier method. The TSDCN is considered to enable high-accuracy classification of high-dimensional time-series patterns and to reduce the computation time taken for network training. The validity of the TSDCN is demonstrated for high-dimensional artificial data and electroencephalogram signals in the experiments conducted during the study.

Continuous fine‐grained arm action recognition using motion spectrum mixture models

Motion sensors in smart wristbands/watches have been widely used to sense users’ level of movement and animation. Some studies have further recognised activity contexts using these sensors, such as walking, sitting and running. However, in applications requiring understanding of more complex activities such as interactions with other people or objects, it is necessary to recognise the fine-grained arm action during user interactions with other people or objects. A method to recognise a set of arm actions on a fine-grained level (e.g. checking the wristband, drinking water etc.) is proposed. Motion signals from the accelerometer and gyroscope are transformed into the frequency domain using the short-time Fourier transform. Then, the action patterns are represented by the motion spectrum mixture model and action dynamics are modelled by continuous density hidden Markov models. Tested on a dataset collected from 23 subjects, the method shows satisfying performance and efficiency.

Test Of Vowels In Speech Recognition Using Continuous Density Hidden Markov Model And Development Of Phonetically Balanced-Words In The Filipino Language

AbstractAn Automatic Speech Recognition (ASR) converts the speech signals into words. The recognized words can be the final output or it can be an input for a natural language processing. In this paper, vowel recognizer using Continuous density HMM and Mel-Frequency Cepstral Coefficient (MFCC) were used for feature extraction for its development, and phonetically balanced words (PBW) in Filipino were developed. Thus, this study is a preparation for Filipino Language ASR using HMM. For vowel recognizer, forty speakers were trained (20 male and 20 female speakers). An average accuracy rate of 94.5% was achieved for speaker-dependent test and 90.8% for speaker independent test. For PBW, 2 word lists were developed consisting of 257 words for the 2-syllable Filipino PBW word list and 212 words for the 3-syllable Filipino PBW word list.

Connected digit speech recognition system for Malayalam language

A connected digit speech recognition is important in many applications such as automated banking system, catalogue-dialing, automatic data entry, automated banking system, etc. This paper presents an optimum speaker-independent connected digit recognizer for Malayalam language. The system employs Perceptual Linear Predictive (PLP) cepstral coefficient for speech parameterization and continuous density Hidden Markov Model (HMM) in the recognition process. Viterbi algorithm is used for decoding. The training data base has the utterance of 21 speakers from the age group of 20 to 40 years and the sound is recorded in the normal office environment where each speaker is asked to read 20 set of continuous digits. The system obtained an accuracy of 99.5 % with the unseen data.

Continuous Density Hidden Markov Model for Hindi Speech Recognition

AbstractState of the art automatic speech recognition system uses Mel frequency cepstral coefficients as feature extractor along with Gaussian mixture model for acoustic modeling but there is no standard value to assign number of mixture component in speech recognition process.Current choice of mixture component is arbitrary with little justification. Also the standard set for European languages can not be used in Hindi speech recognition due to mismatch in database size of the languages.Parameter estimation with too many or few component may inappropriately estimate the mixture model. Therefore, number of mixture is important for initial estimation of expectation maximization process. In this research work, the authors estimate number of Gaussian mixture component for Hindi database based upon the size of vocabulary.Mel frequency cepstral feature and perceptual linear predictive feature along with its extended variations with delta-delta-delta feature have been used to evaluate this number based on optimal recognition score of the system . Comparitive analysis of recognition performance for both the feature extraction methods on medium size Hindi database is also presented in this paper.HLDA has been used as feature reduction technique and also its impact on the recognition score has been highlighted.