Set Of Hidden Markov Models Research Articles

City-scale model for COVID-19 epidemiology with mobility and social activities represented by a set of hidden Markov models

Background and objectiveSARS-CoV-2 emerged by the end of 2019 and became a global pandemic due to its rapid spread. Various outbreaks of the disease in different parts of the world have been studied, and epidemiological analyses of these outbreaks have been useful for developing models with the aim of tracking and predicting the spread of epidemics. In this paper, an agent-based model that predicts the local daily evolution of the number of people hospitalized in intensive care due to COVID-19 is presented. MethodsAn agent-based model has been developed, taking into consideration the most relevant characteristics of the geography and climate of a mid-size city, its population and pathology statistics, and its social customs and mobility, including the state of public transportation. In addition to these inputs, the different phases of isolation and social distancing are also taken into account. By means of a set of hidden Markov models, the system captures and reproduces virus transmission associated with the stochastic nature of people's mobility and activities in the city. The spread of the virus in the host is also simulated by following the stages of the disease and by considering the existence of comorbidities and the proportion of asymptomatic carriers. ResultsAs a case study, the model was applied to Paraná city (Entre Ríos, Argentina) in the second half of 2020. The model adequately predicts the daily evolution of people hospitalized in intensive care due to COVID-19. This adequacy is reflected by the fact that the prediction of the model (including its dispersion), as with the data reported in the field, never exceeded 90% of the capacity of beds installed in the city. In addition, other epidemiological variables of interest, with discrimination by age range, were also adequately reproduced, such as the number of deaths, reported cases, and asymptomatic individuals. ConclusionsThe model can be used to predict the most likely evolution of the number of cases and hospital bed occupancy in the short term. By adjusting the model to match the data on hospitalizations in intensive care units and deaths due to COVID-19, it is possible to analyze the impact of isolation and social distancing measures on the disease spread dynamics. In addition, it allows for simulating combinations of characteristics that would lead to a potential collapse in the health system due to lack of infrastructure as well as predicting the impact of social events or increases in people's mobility.

Diagnosis method based on hidden Markov model and Weibull mixture model for mechanical faults of in-wheel motors

To effectively monitor the operation state of in-wheel motors used in electric vehicles and ensure the safety of the whole vehicle, a diagnosis method based on hidden Markov model (HMM) and Weibull mixture model (WMM) is proposed for mechanical faults in in-wheel motors, known simply as the WMM-HMM diagnosis method. Firstly, vibration signals of the in-wheel motor are extracted for sensitive symptom parameters which are used to characterize the operation state and establish the observation sequence. Secondly, WMM is employed to expand the limited observation sequence under various operating states of in-wheel motors to obtain sufficient observation sequence as the training sample set of HMM, and HMM parameters are determined through combining supervised learning with unsupervised learning algorithm. Then the WMM-HMM diagnosis models are constructed under low and medium speed conditions respectively. Finally, the corresponding faults in-wheel motors are customized and the test bench is built to verify the proposed method. The test results show that the proposed method can accurately identify the mechanical fault state of in-wheel motors under different conditions and has good generalization and applicability in traditional methods comparison.

The Modelling of Movement of Multiple Animals that Share Behavioural Features

AbstractIn this work, we propose a model that can be used to infer the behaviour of multiple animals. Our proposal is defined as a set of hidden Markov models that are based on the sticky hierarchical Dirichlet process, with a shared base-measure, and a step and turn with an attractive point (STAP) emission distribution. The latent classifications are representative of the behaviour assumed by the animals, which is described by the STAP parameters. Given the latent classifications, the animals are independent. As a result of the way we formalize the distribution over the STAP parameters, the animals may share, in different behaviours, the set or a subset of the parameters, thereby allowing us to investigate the similarities between them. The hidden Markov models, based on the Dirichlet process, allow us to estimate the number of latent behaviours for each animal, as a model parameter. This proposal is motivated by a real data problem, where the global positioning system (GPS) coordinates of six Maremma Sheepdogs have been observed. Among the other results, we show that four dogs share most of the behaviour characteristics, while two have specific behaviours.

City-Scale Model for COVID-19 Epidemiology with Mobility and Social Activities Represented by a Set of Hidden Markov Models

City-Scale Model for COVID-19 Epidemiology with Mobility and Social Activities Represented by a Set of Hidden Markov Models

Towards real-time recognition of activities in smart homes

Many supervised methods have been proposed to infer the particular activities of the inhabitants from a variety of sensors attached in the home. Current activity recognition systems either assume that the sensor stream has been presegmented or use a sliding window for activity segmentation. This makes real-time activity recognition task difficult due to the presence of temporal gaps between successive sensor activations. In this paper, we propose a method based on a set of hidden Markov models that can simultaneously solve the problem of activity segmentation and recognition on streaming sensor data without relying on any sliding window methods. We demonstrate our algorithm on sensor data obtained from two publicly available smart homes datasets.

Towards real-time recognition of activities in smart homes

Many supervised methods have been proposed to infer the particular activities of the inhabitants from a variety of sensors attached in the home. Current activity recognition systems either assume that the sensor stream has been presegmented or use a sliding window for activity segmentation. This makes real-time activity recognition task difficult due to the presence of temporal gaps between successive sensor activations. In this paper, we propose a method based on a set of hidden Markov models that can simultaneously solve the problem of activity segmentation and recognition on streaming sensor data without relying on any sliding window methods. We demonstrate our algorithm on sensor data obtained from two publicly available smart homes datasets.

Validating the AMRFinder Tool and Resistance Gene Database by Using Antimicrobial Resistance Genotype-Phenotype Correlations in a Collection of Isolates.

Antimicrobial resistance (AMR) is a major public health problem that requires publicly available tools for rapid analysis. To identify AMR genes in whole-genome sequences, the National Center for Biotechnology Information (NCBI) has produced AMRFinder, a tool that identifies AMR genes using a high-quality curated AMR gene reference database. The Bacterial Antimicrobial Resistance Reference Gene Database consists of up-to-date gene nomenclature, a set of hidden Markov models (HMMs), and a curated protein family hierarchy. Currently, it contains 4,579 antimicrobial resistance proteins and more than 560 HMMs. Here, we describe AMRFinder and its associated database. To assess the predictive ability of AMRFinder, we measured the consistency between predicted AMR genotypes from AMRFinder and resistance phenotypes of 6,242 isolates from the National Antimicrobial Resistance Monitoring System (NARMS). This included 5,425 Salmonella enterica, 770 Campylobacter spp., and 47 Escherichia coli isolates phenotypically tested against various antimicrobial agents. Of 87,679 susceptibility tests performed, 98.4% were consistent with predictions. To assess the accuracy of AMRFinder, we compared its gene symbol output with that of a 2017 version of ResFinder, another publicly available resistance gene detection system. Most gene calls were identical, but there were 1,229 gene symbol differences (8.8%) between them, with differences due to both algorithmic differences and database composition. AMRFinder missed 16 loci that ResFinder found, while ResFinder missed 216 loci that AMRFinder identified. Based on these results, AMRFinder appears to be a highly accurate AMR gene detection system.

Automatic event detection in basketball using HMM with energy based defensive assignment

Abstract We propose a unsupervised learning framework for automatically labeling events in a basketball game. Our framework uses the the optical player tracking data in the NBA. We first learn the time series of defensive assignments using a novel player and location dependent attraction based model which uses hidden Markov models (HMMs), Gaussian processes, and a “bond breaking” model for changes in defensive assignments. Next, we use the learned defensive assignments as an input to a set of HMMs that automatically detect events such as ball screens, drives and post-ups. We show that our models provide significant improvements over existing benchmarks both on defensive assignments and event detection.

Multi-view recognition system for human activity based on multiple features for video surveillance system

Recognition of the activities of human, image sequences is most active area of research in computer vision and most of the previous projects, the focus of the activity is to acknowledge the recognition, from a single view and ignored issues of multiple view invariance. In this paper, the proposed framework for the recognition of a view-invariant human activity, solves the above problem. The components of the proposed framework are three consecutive modules: (i) the detection and positioning of the person’s background subtraction, (ii) the function extraction (iii) and the final activity is referenced by using a set of hidden Markov models (HMMs). During features extraction phase in the proposed method for activity representation a combination of contour-based distance signal feature, optical flow-based motion feature and uniform rotation local binary patterns has been used. Due to its rotation invariant nature the uniform LBP provides view-invariant recognition of multi-view human activities. A successful testing of the proposed approach was done on our own viewpoint dataset, KTH action recognition dataset, i3DPost multi-view dataset, and MSR view-point action dataset. From the experimental results and analysis over the chosen datasets, it is observed that the proposed framework is robust, flexible and efficient with respect to multiple views activity recognition, scale and phase variations.

Sequence Classification Using Third-Order Moments.

Model-based classification of sequence data using a set of hidden Markov models is a well-known technique. The involved score function, which is often based on the class-conditional likelihood, can, however, be computationally demanding, especially for long data sequences. Inspired by recent theoretical advances in spectral learning of hidden Markov models, we propose a score function based on third-order moments. In particular, we propose to use the Kullback-Leibler divergence between theoretical and empirical third-order moments for classification of sequence data with discrete observations. The proposed method provides lower computational complexity at classification time than the usual likelihood-based methods. In order to demonstrate the properties of the proposed method, we perform classification of both simulated data and empirical data from a human activity recognition study.

Video-realistic expressive audio-visual speech synthesis for the Greek language

High quality expressive speech synthesis has been a long-standing goal towards natural human-computer interaction. Generating a talking head which is both realistic and expressive appears to be a considerable challenge, due to both the high complexity in the acoustic and visual streams and the large non-discrete number of emotional states we would like the talking head to be able to express. In order to cover all the desired emotions, a significant amount of data is required, which poses an additional time-consuming data collection challenge. In this paper we attempt to address the aforementioned problems in an audio-visual context. Towards this goal, we propose two deep neural network (DNN) architectures for Video-realistic Expressive Audio-Visual Text-To-Speech synthesis (EAVTTS) and evaluate them by comparing them directly both to traditional hidden Markov model (HMM) based EAVTTS, as well as a concatenative unit selection EAVTTS approach, both on the realism and the expressiveness of the generated talking head. Next, we investigate adaptation and interpolation techniques to address the problem of covering the large emotional space. We use HMM interpolation in order to generate different levels of intensity for an emotion, as well as investigate whether it is possible to generate speech with intermediate speaking styles between two emotions. In addition, we employ HMM adaptation to adapt an HMM-based system to another emotion using only a limited amount of adaptation data from the target emotion. We performed an extensive experimental evaluation on a medium sized audio-visual corpus covering three emotions, namely anger, sadness and happiness, as well as neutral reading style. Our results show that DNN-based models outperform HMMs and unit selection on both the realism and expressiveness of the generated talking heads, while in terms of adaptation we can successfully adapt an audio-visual HMM set trained on a neutral speaking style database to a target emotion. Finally, we show that HMM interpolation can indeed generate different levels of intensity for EAVTTS by interpolating an emotion with the neutral reading style, as well as in some cases, generate audio-visual speech with intermediate expressions between two emotions.

HMM-Based Phrase-Independent i-Vector Extractor for Text-Dependent Speaker Verification

The low-dimensional i-vector representation of speech segments is used in the state-of-the-art text-independent speaker verification systems. However, i-vectors were deemed unsuitable for the text-dependent task, where simpler and older speaker recognition approaches were found more effective. In this work, we propose a straightforward hidden Markov model (HMM) based extension of the i-vector approach, which allows i-vectors to be successfully applied to text-dependent speaker verification. In our approach, the Universal Background Model (UBM) for training phrase-independent i-vector extractor is based on a set of monophone HMMs instead of the standard Gaussian Mixture Model (GMM). To compensate for the channel variability, we propose to precondition i-vectors using a regularized variant of within-class covariance normalization, which can be robustly estimated in a phrase-dependent fashion on the small datasets available for the text-dependent task. The verification scores are cosine similarities between the i-vectors normalized using phrase-dependent s-norm. The experimental results on RSR2015 and RedDots databases confirm the effectiveness of the proposed approach, especially in rejecting test utterances with a wrong phrase. A simple MFCC based i-vector/HMM system performs competitively when compared to very computationally expensive DNN-based approaches or the conventional relevance MAP GMM-UBM, which does not allow for compact speaker representations. To our knowledge, this paper presents the best published results obtained with a single system on both RSR2015 and RedDots dataset.

Hidden Markov Models for the Prediction of Impending Faults

Reliability and safety are two important concepts in industrial applications. Thus, the development of monitoring tools, which are able to ensure the continuity of service by predicting faults, should improve competitiveness. This paper presents two probabilistic methods based on hidden Markov models (HMMs) for the prediction of impending faults. This paper shows that a prediction of faults is not limited to the estimation of the remaining useful life but is also extended to the estimation of the risk of an imminent appearance of faults in the future. The first method consists in modeling the degradation process of the studied system by a single HMM. A probabilistic model is proposed to predict an imminent appearance of a fault. The second method consists in modeling the degradation states by a set of HMMs. Another probabilistic model is proposed to predict an imminent appearance of a fault. An experimental application is proposed to demonstrate their applicability. The obtained results show their effectiveness to predict the imminent appearance of faults.

Action recognition from only somatosensory information using spectral learning in a hidden Markov model

Human action classification is fundamental technology for robots that have to interpret a human’s intended actions and make appropriate responses, as they will have to do if they are to be integrated into our daily lives. Improved measurement of human motion, using an optical motion capture system or a depth sensor, allows robots to recognize human actions from superficial motion data, such as camera images containing human actions or positions of human bodies. But existing technology for motion recognition does not handle the contact force that always exists between the human and the environment that the human is acting upon. More specifically, humans perform feasible actions by controlling not only their posture but also the contact forces. Furthermore these contact forces require appropriate muscle tensions in the full body. These muscle tensions or activities are expected to be useful for robots observing human actions to estimate the human’s somatosensory states and consequently understand the intended action. This paper proposes a novel approach to classifying human actions using only the activities of all the muscles in the human body. Continuous spatio-temporal data of the activity of an individual muscle is encoded into a discrete hidden Markov model (HMM), and the set of HMMs for all the muscles forms a classifier for the specific action. Our classifiers were tested on muscle activities estimated from captured human motions, electromyography data, and reaction forces. The results demonstrate their superiority over commonly used HMM-based classifiers.

Hankelet-based dynamical systems modeling for 3D action recognition

This paper proposes to model an action as the output of a sequence of atomic Linear Time Invariant (LTI) systems. The sequence of LTI systems generating the action is modeled as a Markov chain, where a Hidden Markov Model (HMM) is used to model the transition from one atomic LTI system to another. In turn, the LTI systems are represented in terms of their Hankel matrices. For classification purposes, the parameters of a set of HMMs (one for each action class) are learned via a discriminative approach. This work proposes a novel method to learn the atomic LTI systems from training data, and analyzes in detail the action representation in terms of a sequence of Hankel matrices. Extensive evaluation of the proposed approach on two publicly available datasets demonstrates that the proposed method attains state-of-the-art accuracy in action classification from the 3D locations of body joints (skeleton).

Reliable Accent-Specific Unit Generation With Discriminative Dynamic Gaussian Mixture Selection for Multi-Accent Chinese Speech Recognition

In this paper, we propose a discriminative dynamic Gaussian mixture selection (DGMS) strategy to generate reliable accent-specific units (ASUs) for multi-accent speech recognition. Time-aligned phone recognition is used to generate the ASUs that model accent variations explicitly and accurately. DGMS reconstructs and adjusts a pre-trained set of hidden Markov model (HMM) state densities to build dynamic observation densities for each input speech frame. A discriminative minimum classification error criterion is adopted to optimize the sizes of the HMM state observation densities with a genetic algorithm (GA). To the author's knowledge, the discriminative optimization for DGMS accomplishes discriminative training of discrete variables that is first proposed. We found the proposed framework is able to cover more multi-accent changes, thus reduce some performance loss in pruned beam search, without increasing the model size of the original acoustic model set. Evaluation on three typical Chinese accents, Chuan, Yue and Wu, shows that our approach outperforms traditional acoustic model reconstruction techniques with a syllable error rate reduction of 8.0%, 5.5% and 5.0%, respectively, while maintaining a good performance on standard Putonghua speech.

Bioinformatic prediction of gene functions regulated by quorum sensing in the bioleaching bacterium Acidithiobacillus ferrooxidans.

The biomining bacterium Acidithiobacillus ferrooxidans oxidizes sulfide ores and promotes metal solubilization. The efficiency of this process depends on the attachment of cells to surfaces, a process regulated by quorum sensing (QS) cell-to-cell signalling in many Gram-negative bacteria. At. ferrooxidans has a functional QS system and the presence of AHLs enhances its attachment to pyrite. However, direct targets of the QS transcription factor AfeR remain unknown. In this study, a bioinformatic approach was used to infer possible AfeR direct targets based on the particular palindromic features of the AfeR binding site. A set of Hidden Markov Models designed to maintain palindromic regions and vary non-palindromic regions was used to screen for putative binding sites. By annotating the context of each predicted binding site (PBS), we classified them according to their positional coherence relative to other putative genomic structures such as start codons, RNA polymerase promoter elements and intergenic regions. We further used the Multiple EM for Motif Elicitation algorithm (MEME) to further filter out low homology PBSs. In summary, 75 target-genes were identified, 34 of which have a higher confidence level. Among the identified genes, we found afeR itself, zwf, genes encoding glycosyltransferase activities, metallo-beta lactamases, and active transport-related proteins. Glycosyltransferases and Zwf (Glucose 6-phosphate-1-dehydrogenase) might be directly involved in polysaccharide biosynthesis and attachment to minerals by At. ferrooxidans cells during the bioleaching process.

Compensation of SNR and noise type mismatch using an environmental sniffing based speech recognition solution

Multiple-model based speech recognition (MMSR) has been shown to be quite successful in noisy speech recognition. Since it employs multiple hidden Markov model (HMM) sets that correspond to various noise types and signal-to-noise ratio (SNR) values, the selected acoustic model can be closely matched with the test noisy speech, which leads to improved performance when compared with other state-of-the-art speech recognition systems that employ a single HMM set. However, as the number of HMM sets is usually limited due to practical considerations as well as effective model selection, acoustic mismatch can still be a problem in MMSR. In this study, we proposed methods to improve recognition performance by mitigating the mismatch in SNR and noise type for an MMSR solution. For the SNR mismatch, an optimal SNR mapping between the test noisy speech and the HMM was determined by experimental investigation. Improved performance was demonstrated by employing the SNR mapping instead of using the estimated SNR of the test noisy speech directly. We also proposed a novel method to reduce the effect of noise type mismatch by compensating the test noisy speech in the log-spectrum domain. We first derive the relation between the log-spectrum vectors in the test and training noisy speech. Since the relation is a non-linear function of the speech and noise parameters, the statistical information regarding the testing log-spectrum vectors was obtained by approximation using vector Taylor series (VTS) algorithm. Finally, the minimum mean square error estimation of the training log-spectrum vectors was used to reduce the mismatch between the training and test noisy speech. By employing the proposed methods in the MMSR framework, relative word error rate reduction of 18.7% and 21.3% was achieved on the Aurora 2 task when compared to a conventional MMSR and multi-condition training (MTR) method, respectively.

Audiovisual Voice Activity Detection Based on Microphone Arrays and Color Information

Audiovisual voice activity detection is a necessary stage in several problems, such as advanced teleconferencing, speech recognition, and human-computer interaction. Lip motion and audio analysis provide a large amount of information that can be integrated to produce more robust audiovisual voice activity detection (VAD) schemes, as we discuss in this paper. Lip motion is very useful for detecting the active speaker, and in this paper we introduce a new approach for lips and visual VAD. First, the algorithm performs skin segmentation to reduce the search area for lip extraction, and the most likely lip and non-lip regions are detected using a Bayesian approach within the delimited area. Lip motion is then detected using Hidden Markov Models (HMMs) that estimate the likely occurrence of active speech within a temporal window. Audio information is captured by an array of microphones, and the sound-based VAD is related to finding spatio-temporally coherent sound sources through another set of HMMs. To increase the robustness of the proposed system, a late fusion approach is employed to combine the result of each modality (audio and video). Our experimental results indicate that the proposed audiovisual approach presents better results when compared to existing VAD algorithms.

Speaker-dependent model interpolation for statistical emotional speech synthesis

In this article, we propose a speaker-dependent model interpolation method for statistical emotional speech synthesis. The basic idea is to combine the neutral model set of the target speaker and an emotional model set selected from a pool of speakers. For model selection and interpolation weight determination, we propose to use a novel monophone-based Mahalanobis distance, which is a proper distance measure between two Hidden Markov Model sets. We design Latin-square evaluation to reduce the systematic bias in the subjective listening tests. The proposed interpolation method achieves sound performance on the emotional expressiveness, the naturalness, and the target speaker similarity. Moreover, such performance is achieved without the need to collect the emotional speech of the target speaker, saving the cost of data collection and labeling.