结合HMM和SPSM的手语识别方法

手语识别的研究具有重大的学术价值和广泛的应用前景。在近些年的手语识别工作中,隐马尔可夫模型(Hidden Markov Models,简称HMMs)起到了重要的作用,但是,HMMs假设同一状态内的观察值之间是独立同分布的,这个假设同某些手语信号的帧间相关性相背离。受到多项式片段模型(Polynomial Segment Models,简称PSMs)能够显式描述帧间相关性的启发,提出了一种简化的PSMs,其中应用马氏距离作为距离测度。实验表明,这种简化的PSMs在同传统的HMMs进行后验概率归一化求和的融合之后,手语词的平均相对正确率得到了13.38%的提升,从而证明此方法是一种更加精确的手语识别方法。     

基于高斯混合模型的语音性别识别

利用高斯混合模型(GMM)方法进行语音的性别识别.首先概述了特征提取、识别方法及性别识别的过程;然后通过减少提取特征的语音帧数和降低高斯混合模型的混合阶数来提高性别识别速度;最后,将由Mel频率倒谱参数(MFCC)特征和基音频率特征两种方法得到的测试样本后验概率结合,提出新的计算测试样本后验概率的方法.实验表明依据此后验概率能有效提高识别的正确率.     

支持向量机中引入后验概率的理论和方法研究

目前支持向量机解决模式识别问题是广大学者研究的热点，样本的后验概率在模式识别中至关重要，但是传统的支持向量机技术不提供后验概率，针对这一问题进行了3个方面的研究：（1）在给出样本点后验概率的基础上，将大规模优化问题分解成最大似然函数和最大分类边界两个规模优化问题；（2）给出了一种新的用后验概率修正最优分离超平面的方法，并且分析了该新方法的合理性；（3）用图像分类的3组实例说明本方法的有效性。     

加权抽样对相似性学习算法的改进效果研究

当今诸多聚类算法需要通过计算样本间距离来得到样本相似性。因此对这类算法而言,距离的计算方法尤为重要。对部分现有距离度量学习或相似性学习算法进行研究后可以发现,多数算法在选择学习样本的过程中,都采用了重复随机抽样的方式。这一抽样方式使所有训练节点都有均等概率用于度量或相似性学习,但因样本位置不同,对分类算法而言样本的分类难度也不同。如果能针对较难分类的样本进行着重学习,并适当减少对易分类点的学习时间,便能提高学习过程的效率性,减少学习过程的时间。节约时间成本,在大数据时代有不容忽视的意义。     

基于可见光与红外数据融合的地形分类

针对单传感器地形分类效果不佳的问题,提出一种基于可见光与红外数据融合的地形分类方法。分别对可见光图像与红外图像提取特征,使用最近邻分类器和最小距离分类器进行后验概率估计,将来自不同特征、不同分类器的后验概率加权组合,通过散度计算得到特征的权重,实验确定分类器的权重,并在最小距离的后验概率估计中,使用马氏距离代替欧氏距离。实验结果表明,该方法对水泥路和沙子路的识别率分别达到99.33%和96.67%,均高于同类方法。     

后验概率支持向量机在企业信用评级中的应用

在支持向量机(Support Vector Machine)的分类问题中,训练样本的分类信息总是确定的,由此得到的分类指示函数也总是对新样本给出确定的分类信息,但是这种情况对一些不确定性问题并不恰当.利用贝叶斯规则,将样本的后验概率与传统支持向量机结合,得到了基于后验概率的支持向量机.在具体的算法上,引入了一个经验性的方法得到样本的后验概率.以某评级机构提供的企业信用评估数据库为研究对象.     

后验概率在多分类支持向量机上的应用

支持向量机是基于统计学习理论的一种新的分类规则挖掘方法。在已有多分类支持向量机基础上,首次提出了几何距离多分类支持向量分类器;随后,将二值支持向量机的后验概率输出也推广到多分类问题,避免了使用迭代算法,在快速预测的前提下提高了预测准确率。数值实验的结果表明,这两种方法都具有很好的推广性能,能明显提高分类器对未知样本的分类准确率。     

多预测子融合实时连续语音识别输出词正误判别

本文在采用堆栈译码词网重估输出作为识别最终输出的连续语音识别实时解码条件下,利用决策树方法将多个预测子融合,对识别输出词进行正确和错误的判别。本文首先构造了词后验概率、词长、相邻词的后验概率、词的声学和语言得分等共13 个预测子,然后利用决策树方法,通过选择不同的预测子组合方式和适当的决策树建树参数,筛选出预测子的最佳组合,建立优化的决策树进行输出词的正误判别。实验结果表明:利用局域词图计算的词后验概率与词长、相邻词的后验概率等几种实时预测子融合后,对识别输出词的正误判别能力得到提高,并且在实时性和分类效果两个方面优于n - best 输出的相应结果,相对于基线系统, 则分类错误率下降41. 4 %。实验结果也表明本文提出的相邻词的后验概率是相对重要的预测子。     

基于节点生长马氏距离K均值和HMM的网络入侵检测方法设计

传统的网络入侵检测方法存在着检测率低和无法进行在线检测的问题,为此设计了一种基于节点生长马氏距离K均值和HMM的网络入侵检测方法;首先,给出了入侵检测系统框图,然后,以马氏距离为评价准则,提出了一种节点根据距离阈值进行自适应生长的K均值算法以实现样本的聚类,得到样本属于各攻击类型的后验概率,并采用此后验概率来初始化HMM中的初始矢量分布、状态转移概率和观察值概率等参数,通过前向评估准则和后向评估准则对HMM模型进行训练,从而获得了HMM检测模型,将样本输入到各检测模型中并将概率最大的检测模型作为其攻击类型;仿真试验表明所提方法能有效地实现网络入侵检测,不仅具有较高的检测率,而且具有较低的误检率和漏检率,是一种有效的网络入侵检测方法。     

一种基于反例样本修剪支持向量机的事件追踪算法

支持向量机（SVM）在各类别样本数目分布不均匀时，样本数量越多其分类误差越小，而样本数量越少其分类误差越大．在分析这种倾向产生原因的基础上，提出了一种基于反例样本修剪支持向量机（NEP—SVM）的事件追踪算法．该算法首先修剪反例样本，根据距离和类标决定一反例样本的取舍，然后使用SVM对新的样本集进行训练以得到分类器，补偿了上述倾向性问题造成的不利影响．另外，由于后验概率对于提高事件追踪的性能至关重要，而传统的支持向量机不提供后验概率，本文通过一个sigmoid函数的参数训练将SVM的输出结果映射成概率．实验结果表明NEP—SVM是有效的．     

Average divergence distance as a statistical discrimination measure for hidden Markov models

This paper proposes and evaluates a new statistical discrimination measure for hidden Markov models (HMMs) extending the notion of divergence, a measure of average discrimination information originally defined for two probability density functions. Similar distance measures have been proposed for the case of HMMs, but those have focused primarily on the stationary behavior of the models. However, in speech recognition applications, the transient aspects of the models have a principal role in the discrimination process and, consequently, capturing this information is crucial in the formulation of any discrimination indicator. This paper proposes the notion of average divergence distance (ADD) as a statistical discrimination measure between two HMMs, considering the transient behavior of these models. This paper provides an analytical formulation of the proposed discrimination measure, a justification of its definition based on the Viterbi decoding approach, and a formal proof that this quantity is well defined for a left-to-right HMM topology with a final nonemitting state, a standard model for basic acoustic units in automatic speech recognition (ASR) systems. Using experiments based on this discrimination measure, it is shown that ADD provides a coherent way to evaluate the discrimination dissimilarity between acoustic models.     

Gene clustering with hidden Markov model optimized by PSO algorithm

Gene clustering is one of the most important problems in bioinformatics. In the sequential data clustering, hidden Markov models (HMMs) have been widely used to find similarity between sequences, due to their capability of handling sequence patterns with various lengths. In this paper, a novel gene clustering scheme based on HMMs optimized by particle swarm optimization algorithm is introduced. In this approach, each gene sequence is described by a specific HMM, and then for each model, its probability to generate individual sequence is evaluated. A hierarchical clustering algorithm based on a new definition of a distance measure has been applied to find the best clusters. Experiments carried out on lung cancer-related genes dataset show that the proposed approach can be successfully utilized for gene clustering.     

Vague（值）集间的相似度量及其应用

鉴于已知的一些Vague集间的相似度量和距离公式有缺陷,提出用分段函数表达的Vague（值）集间的接近度的定义,应用此定义重新给出用分段函数表达的Vague（值）集相似度量的定义。给出了三个加权接近度公式和三个加权相似度量新公式。给出在Vague环境下用Vague集间的接近度和相似度量进行模式识别的方法。应用实例表明所给公式皆是有效的。     

Most discriminating segment – Longest common subsequence (MDSLCS) algorithm for dynamic hand gesture classification

In this work, we consider the recognition of dynamic gestures based on representative sub-segments of a gesture, which are denoted as most discriminating segments (MDSs). The automatic extraction and recognition of such small representative segments, rather than extracting and recognizing the full gestures themselves, allows for a more discriminative classifier. A MDS is a sub-segment of a gesture that is most dissimilar to all other gesture sub-segments. Gestures are classified using a MDSLCS algorithm, which recognizes the MDSs using a modified longest common subsequence (LCS) measure. The extraction of MDSs from a data stream uses adaptive window parameters, which are driven by the successive results of multiple calls to the LCS classifier. In a preprocessing stage, gestures that have large motion variations are replaced by several forms of lesser variation. We learn these forms by adaptive clustering of a training set of gestures, where we reemploy the LCS to determine similarity between gesture trajectories. The MDSLCS classifier achieved a gesture recognition rate of 92.6% when tested using a set of pre-cut free hand digit (0–9) gestures, while hidden Markov models (HMMs) achieved an accuracy of 89.5%. When the MDSLCS was tested against a set of streamed digit gestures, an accuracy of 89.6% was obtained. At present the HMMs method is considered the state-of-the-art method for classifying motion trajectories. The MDSLCS algorithm had a higher accuracy rate for pre-cut gestures, and is also more suitable for streamed gestures. MDSLCS provides a significant advantage over HMMs by not requiring data re-sampling during run-time and performing well with small training sets.     

Probabilistic distances between finite-state finite-alphabet hidden Markov models

This note considers the problem of evaluating a probabilistic distance between discrete-time, homogeneous, first-order, finite-state finite-alphabet hidden Markov models (HMMs). Our approach is based on a correspondence between probability measures and HMMs established in this note. Using a probability measure transformation technique, we obtain recursive expressions for the relative entropy between the marginal probability distributions of two HMMs under consideration. Also, the relative entropy rate, as the limit of the time-averaged value of the above relative entropy, is obtained. These expressions are given in terms of the parameters of the given HMMs. Furthermore, we show that the probabilistic distance between HMMs used in the existing literature admits a representation in terms of a conditional expectation given the observation sequence. This representation allows us to evaluate this distance using an information state approach.     

Hybrid continuous speech recognition systems by HMM,MLP and SVM: a comparative study

This paper presents a new hybrid method for continuous Arabic speech recognition based on triphones modelling. To do this, we apply Support Vectors Machine (SVM) as an estimator of posterior probabilities within the Hidden Markov Models (HMM) standards. In this work, we describe a new approach of categorising Arabic vowels to long and short vowels to be applied on the labeling phase of speech signals. Using this new labeling method, we deduce that SVM/HMM hybrid model is more efficient then HMMs standards and the hybrid system Multi-Layer Perceptron (MLP) with HMM. The obtained results for the Arabic speech recognition system based on triphones are 64.68 % with HMMs, 72.39 % with MLP/HMM and 74.01 % for SVM/HMM hybrid model. The WER obtained for the recognition of continuous speech by the three systems proves the performance of SVM/HMM by obtaining the lowest average for 4 tested speakers 11.42 %.     

Line segment Hausdorff distance on face matching

A novel concept of line segment Hausdorff distance is proposed in this paper. Researchers apply Hausdorff distance to measure the similarity of two point sets. It is extended here to match two sets of line segments. The new approach has the advantage to incorporate structural and spatial information to compute the similarity. The added information can conceptually provide more and better distinctive capability for recognition. This would strengthen and enhance the matching process of similar objects such as faces. The proposed technique has been applied online segments generated from the edge maps of faces with encouraging result that supports the concept experimentally. The results also implicate that line segments could provide sufficient information for face recognition. This might imply a new way for face coding and recognition.     

Discriminant analysis and similarity measure

The cosine similarity measure is often applied after discriminant analysis in pattern recognition. This paper first analyzes why the cosine similarity is preferred by establishing the connection between the cosine similarity based decision rule in the discriminant analysis framework and the Bayes decision rule for minimum error. The paper then investigates the challenges inherent of the cosine similarity and presents a new similarity that overcomes these challenges. The contributions of the paper are thus three-fold. First, the application of the cosine similarity after discriminant analysis is discovered to have its theoretical roots in the Bayes decision rule. Second, some inherent problems of the cosine similarity such as its inadequacy in addressing distance and angular measures are discussed. Finally, a new similarity measure, which overcomes the problems by integrating the absolute value of the angular measure and the l_p norm (the distance measure), is presented to enhance pattern recognition performance. The effectiveness of the proposed new similarity measure in the discriminant analysis framework is evaluated using a large scale, grand challenge problem, namely, the Face Recognition Grand Challenge (FRGC) problem. Experimental results using 36,818 FRGC images on the most challenging FRGC experiment, the FRGC Experiment 4, show that the new similarity measure improves face recognition performance upon other popular similarity measures, such as the cosine similarity measure, the normalized correlation, and the Euclidean distance measure.     

Texture recognition via autoregression

This article considers the image processing problem of texture recognition. It is shown that a chi-square test based upon a two-dimensional autoregressive model can be derived and can be used to test for differences between certain types of micro-textures. The chi-square test cannot be used with macro-textures, and another autoregressive-based distance measure is derived which is more suitable for these cases. It is shown experimentally that this distance measure affords a reliable means of classifying a broad class of micro- and macro-textures using a nearest-neighbour type of approach.