Similarity-based classification of sequences using hidden Markov models

Hidden Markov models (HMM) are a widely used tool for sequence modelling. In the sequence classification case, the standard approach consists of training one HMM for each class and then using a standard Bayesian classification rule. In this paper, we introduce a novel classification scheme for sequences based on HMMs, which is obtained by extending the recently proposed similarity-based classification paradigm to HMM-based classification. In this approach, each object is described by the vector of its similarities with respect to a predetermined set of other objects, where these similarities are supported by HMMs. A central problem is the high dimensionality of resulting space, and, to deal with it, three alternatives are investigated. Synthetic and real experiments show that the similarity-based approach outperforms standard HMM classification schemes.     

HMM-based handwritten word recognition: on the optimization of the number of states, training iterations and Gaussian components

In off-line handwriting recognition, classifiers based on hidden Markov models (HMMs) have become very popular. However, while there exist well-established training algorithms which optimize the transition and output probabilities of a given HMM architecture, the architecture itself, and in particular the number of states, must be chosen “by hand”. Also the number of training iterations and the output distributions need to be defined by the system designer. In this paper we examine several optimization strategies for an HMM classifier that works with continuous feature values. The proposed optimization strategies are evaluated in the context of a handwritten word recognition task.     

Iterative Boolean combination of classifiers in the ROC space: An application to anomaly detection with HMMs

Hidden Markov models (HMMs) have been shown to provide a high level performance for detecting anomalies in sequences of system calls to the operating system kernel. Using Boolean conjunction and disjunction functions to combine the responses of multiple HMMs in the ROC space may significantly improve performance over a “single best” HMM. However, these techniques assume that the classifiers are conditional independent, and their of ROC curves are convex. These assumptions are violated in most real-world applications, especially when classifiers are designed using limited and imbalanced training data. In this paper, the iterative Boolean combination (IBC) technique is proposed for efficient fusion of the responses from multiple classifiers in the ROC space. It applies all Boolean functions to combine the ROC curves corresponding to multiple classifiers, requires no prior assumptions, and its time complexity is linear with the number of classifiers. The results of computer simulations conducted on both synthetic and real-world host-based intrusion detection data indicate that the IBC of responses from multiple HMMs can achieve a significantly higher level of performance than the Boolean conjunction and disjunction combinations, especially when training data are limited and imbalanced. The proposed IBC is general in that it can be employed to combine diverse responses of any crisp or soft one- or two-class classifiers, and for wide range of application domains.     

Maximum entropy direct models for speech recognition

Traditional statistical models for speech recognition have mostly been based on a Bayesian framework using generative models such as hidden Markov models (HMMs). This paper focuses on a new framework for speech recognition using maximum entropy direct modeling, where the probability of a state or word sequence given an observation sequence is computed directly from the model. In contrast to HMMs, features can be asynchronous and overlapping. This model therefore allows for the potential combination of many different types of features, which need not be statistically independent of each other. In this paper, a specific kind of direct model, the maximum entropy Markov model (MEMM), is studied. Even with conventional acoustic features, the approach already shows promising results for phone level decoding. The MEMM significantly outperforms traditional HMMs in word error rate when used as stand-alone acoustic models. Preliminary results combining the MEMM scores with HMM and language model scores show modest improvements over the best HMM speech recognizer.     

Ensemble of HMM classifiers based on the clustering validity index for a handwritten numeral recognizer

A new scheme for the optimization of codebook sizes for Hidden Markov Models (HMMs) and the generation of HMM ensembles is proposed in this paper. In a discrete HMM, the vector quantization procedure and the generated codebook are associated with performance degradation. By using a selected clustering validity index, we show that the optimization of HMM codebook size can be selected without training HMM classifiers. Moreover, the proposed scheme yields multiple optimized HMM classifiers, and each individual HMM is based on a different codebook size. By using these to construct an ensemble of HMM classifiers, this scheme can compensate for the degradation of a discrete HMM.

Diagnostic tools for evaluating and updating hidden Markov models

In this paper we consider two related problems in hidden Markov models (HMMs). One, how the various parameters of an HMM actually contribute to predictions of state sequences and spatio-temporal pattern recognition. Two, how the HMM parameters (and associated HMM topology) can be updated to improve performance. These issues are examined in the context of four different experimental settings from pure simulations to observed data. Results clearly demonstrate the benefits of applying some critical tests on the model parameters before using it as a predictor or spatio-temporal pattern recognition technique.     

Large vocabulary word recognition using context-dependent allophonic hidden Markov models

In this paper, we report our development of context-dependent allophonic hidden Markov models (HMMs) implemented in a 75 000-word speaker-dependent Gaussian-HMM recognizer. The context explored is the immediate left and/or right adjacent phoneme. To achieve reliable estimation of the model parameters, phonemes are grouped into classes based on their expected co-articulatory effects on neighboring phonemes. Only five separate preceding and following contexts are identified explicitly for each phoneme. By grouping the contexts we ensure that they occur frequently enough in the training data to allow reliable estimation of the parameters of the HMM representing the context-dependent units. Further improvement in the estimation reliability is obtained by tying the covariance matrices in the HMM output distributions across all contexts. Speech recognition experiments show that when a large amount of data (e.g. over 2500 words) is used to train context-dependent HMMs, the word recognition error rate is reduced by 33%, compared with the context-independent HMMs. For smaller amounts of training data the error reduction becomes less significant.     

Hidden Markov models with factored Gaussian mixtures densities

We present a factorial representation of Gaussian mixture models for observation densities in hidden Markov models (HMMs), which uses the factorial learning in the HMM framework. We derive the reestimation formulas for estimating the factorized parameters by the Expectation Maximization (EM) algorithm and propose a novel method for initializing them. To compare the performances of the proposed models with that of the factorial hidden Markov models and HMMs, we have carried out extensive experiments which show that this modelling approach is effective and robust.     

Factorial Hidden Markov Models

Hidden Markov models (HMMs) have proven to be one of the most widely used tools for learning probabilistic models of time series data. In an HMM, information about the past is conveyed through a single discrete variable—the hidden state. We discuss a generalization of HMMs in which this state is factored into multiple state variables and is therefore represented in a distributed manner. We describe an exact algorithm for inferring the posterior probabilities of the hidden state variables given the observations, and relate it to the forward–backward algorithm for HMMs and to algorithms for more general graphical models. Due to the combinatorial nature of the hidden state representation, this exact algorithm is intractable. As in other intractable systems, approximate inference can be carried out using Gibbs sampling or variational methods. Within the variational framework, we present a structured approximation in which the the state variables are decoupled, yielding a tractable algorithm for learning the parameters of the model. Empirical comparisons suggest that these approximations are efficient and provide accurate alternatives to the exact methods. Finally, we use the structured approximation to model Bach's chorales and show that factorial HMMs can capture statistical structure in this data set which an unconstrained HMM cannot.     

Recognition of degraded characters using dynamic Bayesian networks

In this paper, we investigate the application of dynamic Bayesian networks (DBNs) to the recognition of degraded characters. DBNs are an extension of one-dimensional hidden Markov models (HMMs) which can handle several observation and state sequences. In our study, characters are represented by the coupling of two HMM architectures into a single DBN model. The interacting HMMs are a vertical HMM and a horizontal HMM whose observable outputs are the image columns and image rows, respectively. Various couplings are proposed where interactions are achieved through the causal influence between state variables. We compare non-coupled and coupled models on two tasks: the recognition of artificially degraded handwritten digits and the recognition of real degraded old printed characters. Our models show that coupled architectures perform more accurately on degraded characters than basic HMMs, the linear combination of independent HMM scores, as well as discriminative methods such as support vector machines (SVMs).     

基于HMM方法的银行票据自动识别

利用隐态马尔可夫模型(HMMs)，对银行票据中金额的大小写数据识别问题进行了研究．主要内容包括建立新颖的文字分刻算法；设计HMM训练和识别算法．在HMM系统中，将使用频率比较高的手写体错别字和同音字作为不同的字符类来处理；同时在HMM的训练过程中，提出了平滑参数的新方法．实验结果表明，该方法在实践中是可行的，在银行票据自动识别中有很好的应用前景．     

The most probable annotation problem in HMMs and its application to bioinformatics

Hidden Markov models (HMMs) are often used for biological sequence annotation. Each sequence feature is represented by a collection of states with the same label. In annotating a new sequence, we seek the sequence of labels that has highest probability. Computing this most probable annotation was shown NP-hard by Lyngsø and Pedersen [R.B. Lyngsø, C.N.S. Pedersen, The consensus string problem and the complexity of comparing hidden Markov models, J. Comput. System Sci. 65 (3) (2002) 545–569]. We improve their result by showing that the problem is NP-hard for a specific HMM, and present efficient algorithms to compute the most probable annotation for a large class of HMMs, including abstractions of models previously used for transmembrane protein topology prediction and coding region detection. We also present a small experiment showing that the maximum probability annotation is more accurate than the labeling that results from simpler heuristics.     

Development of head detection and tracking systems for visual surveillance

This paper proposes a technique for the detection of head nod and shake gestures based on eye tracking and head motion decision. The eye tracking step is divided into face detection and eye location. Here, we apply a motion segmentation algorithm that examines differences in moving people’s faces. This system utilizes a Hidden Markov Model-based head detection module that carries out complete detection in the input images, followed by the eye tracking module that refines the search based on a candidate list provided by the preprocessing module. The novelty of this paper is derived from differences in real-time input images, preprocessing to remove noises (morphological operators and so on), detecting edge lines and restoration, finding the face area, and cutting the head candidate. Moreover, we adopt a K-means algorithm for finding the head region. Real-time eye tracking extracts the location of eyes from the detected face region and is performed at close to a pair of eyes. After eye tracking, the coordinates of the detected eyes are transformed into a normalized vector of x-coordinate and y-coordinate. Head nod and shake detector uses three hidden Markov models (HMMs). HMM representation of the head detection can estimate the underlying HMM states from a sequence of face images. Head nod and shake can be detected by three HMMs that are adapted by a directional vector. The directional vector represents the direction of the head movement. The vector is HMMs for determining neutral as well as head nod and shake. These techniques are implemented on images, and notable success is notified.     

Classification of process trends based on fuzzified symbolic representation and hidden Markov models

This paper presents a strategy to represent and classify process data for detection of abnormal operating conditions. In representing the data, a wavelet-based smoothing algorithm is used to filter the high frequency noise. A shape analysis technique called triangular episodes then converts the smoothed data into a semi-qualitative form. Two membership functions are implemented to transform the quantitative information in the triangular episodes to a purely symbolic representation. The symbolic data is classified with a set of sequence matching hidden Markov models (HMMs), and the classification is improved by utilizing a time correlated HMM after the sequence matching HMM. The method is tested on simulations with a non-isothermal CSTR and compared with methods that use a back-propagation neural network with and without an ARX model.     

Large margin hidden Markov models for speech recognition

In this paper, motivated by large margin classifiers in machine learning, we propose a novel method to estimate continuous-density hidden Markov model (CDHMM) for speech recognition according to the principle of maximizing the minimum multiclass separation margin. The approach is named large margin HMM. First, we show this type of large margin HMM estimation problem can be formulated as a constrained minimax optimization problem. Second, we propose to solve this constrained minimax optimization problem by using a penalized gradient descent algorithm, where the original objective function, i.e., minimum margin, is approximated by a differentiable function and the constraints are cast as penalty terms in the objective function. The new training method is evaluated in the speaker-independent isolated E-set recognition and the TIDIGITS connected digit string recognition tasks. Experimental results clearly show that the large margin HMMs consistently outperform the conventional HMM training methods. It has been consistently observed that the large margin training method yields significant recognition error rate reduction even on top of some popular discriminative training methods.     

Embedding HMMs-based models in a Euclidean space: the topological hidden Markov models

Current extensions of hidden Markov models such as structural, hierarchical, coupled, and others have the power to classify complex and highly organized patterns. However, one of their major limitations is the inability to cope with topology: When applied to a visible observation (VO) sequence, the traditional HMM-based techniques have difficulty predicting the n-dimensional shape formed by the symbols of the VO sequence. To fulfill this need, we propose a novel paradigm named “topological hidden Markov models” (THMMs) that classifies VO sequences by embedding the nodes of an HMM state transition graph in a Euclidean space. This is achieved by modeling the noise embedded in the shape generated by the VO sequence. We cover the first and second level topological HMMs. We describe five basic problems that are assigned to a second level topological hidden Markov model: (1) sequence probability evaluation, (2) statistical decoding, (3) structural decoding, (4) topological decoding, and (5) learning. To show the significance of this research, we have applied the concept of THMMs to: (i) predict the ASCII class assigned to a handwritten numeral, and (ii) map protein primary structures to their 3D folds. The results show that the second level THMMs outperform the SHMMs and the multi-class SVM classifiers significantly.     

HMM/SVM segmentation and labelling of Arabic speech for speech recognition applications

Building a large vocabulary continuous speech recognition (LVCSR) system requires a lot of hours of segmented and labelled speech data. Arabic language, as many other low-resourced languages, lacks such data, but the use of automatic segmentation proved to be a good alternative to make these resources available. In this paper, we suggest the combination of hidden Markov models (HMMs) and support vector machines (SVMs) to segment and to label the speech waveform into phoneme units. HMMs generate the sequence of phonemes and their frontiers; the SVM refines the frontiers and corrects the labels. The obtained segmented and labelled units may serve as a training set for speech recognition applications. The HMM/SVM segmentation algorithm is assessed using both the hit rate and the word error rate (WER); the resulting scores were compared to those provided by the manual segmentation and to those provided by the well-known embedded learning algorithm. The results show that the speech recognizer built upon the HMM/SVM segmentation outperforms in terms of WER the one built upon the embedded learning segmentation of about 0.05%, even in noisy background.     

Zero knowledge hidden Markov model inference

Hidden Markov models (HMMs) are widely used in pattern recognition. HMM construction requires an initial model structure that is used as a starting point to estimate the model’s parameters. To construct a HMM without a priori knowledge of the structure, we use an approach developed by Crutchfield and Shalizi that requires only a sequence of observations and a maximum data window size. Values of the maximum data window size that are too small result in incorrect models being constructed. Values that are too large reduce the number of data samples that can be considered and exponentially increase the algorithm’s computational complexity. In this paper, we present a method for automatically inferring this parameter directly from training data as part of the model construction process. We present theoretical and experimental results that confirm the utility of the proposed extension.     

Analysis of 3D Hand Trajectory Gestures Using Stroke-Based Composite Hidden Markov Models

We present a glove-based hand gesture recognition system using hidden Markov models (HMMs) for recognizing the unconstrained 3D trajectory gestures of operators in a remote work environment. A Polhemus sensor attached to a PinchGlove is employed to obtain a sequence of 3D positions of a hand trajectory. The direct use of 3D data provides more naturalness in generating gestures, thereby avoiding some of the constraints usually imposed to prevent performance degradation when trajectory data are projected into a specific 2D plane. We use two kinds of HMMs according to the basic units to be modeled: gesture-based HMM and stroke-based HMM. The decomposition of gestures into more primitive strokes is quite attractive, since reversely concatenating stroke-based HMMs makes it possible to construct a new set of gesture-based HMMs. Any deterioration in performance and reliability arising from decomposition can be remedied by a fine-tuned relearning process for such composite HMMs. We also propose an efficient method of estimating a variable threshold of reliability for an HMM, which is found to be useful in rejecting unreliable patterns. In recognition experiments on 16 types of gestures defined for remote work, the fine-tuned composite HMM achieves the best performance of 96.88% recognition rate and also the highest reliability.