采用原始结构的上下文知识对手写体数字串进行分割和识别

首先,本文提出了采用原始结构的上下文知识对非限制手写体数字串进行分割和识别。开发的新算法在数字串图像中确定特征点,以产生可能的分割假设。一种原始的识别图表利用分割假设的空间。分割假设的估算采用新颖的评价图表,以改善系统的分离物阻割。文中的原始算法试图通过搜索来获取分割假设的总数,并寻找最高的分割／识别可信度。NIST NSTRING SD19和CENPARMI数据库用作估算本方法。实验显示,在分割中采用适当的上下文知识可以极大地改善系统的特性。在手写体数字串中,采用神经网络和支持向量机分类器,我们的系统可分别获取95．28％和96．42％的正确识别率。     

基于自适应双阈值的地下目标自动检测算法

在使用探地雷达（GPR）生成的Bscan图像进行地下目标检测时，当前基于深度学习的目标检测网络模型存在训练样本需求量高、耗时长，不能区分目标显著程度，难以识别复杂目标等问题。针对以上问题，提出一种基于直方图的双阈值分割算法。首先，根据地下目标的GPR图像直方图分布特性，快速从直方图中计算出分割地下目标所需的两个阈值；然后，采用支持向量机（SVM）和LeNet的组合分类器模型对分割结果进行分类识别；最后，进行分类结果整合并统计精确度数值。相较于传统的最大类间方差法（Ostu）、迭代法等阈值分割算法，所提算法获得的地下目标分割结果结构更加完整，并且几乎不含噪声。在真实数据集上，所提算法的平均识别准确率达到了90%以上，比仅使用单一分类器的平均识别准确率提高40%以上。实验结果表明，所提算法能够同时有效分割显著和非显著性地下目标，且采用的组合分类器能够获得更好的分类结果，适用于小样本数据集的地下目标自动检测和识别。     

联机手写维文字符切分算法研究

随着手机短信业务普及,智能手机中实现维吾尔文输入、输出已经是新疆地区1000多万少数民族用户迫切的需求。在连续输入的维吾尔文文章或单词中,切分出一个个的字母,供后续的字母识别使用,字母切分是手写输入识别的核心关键技术。手写维文字符串的分割与字符识别密切相关。采用基于识别的分割方法,系统先通过粗略的图像分析寻找所有可能的切点,在分割的过程中引入识别机制来识别分割碎片,将识别结果经过差值运算后置为每个识别对象的识别可信度,利用移动窗口法找到最佳分割路径。在分类器训练时,采用特征提取来估计分类器参数,得到了性质良好的分类器,试验表明,字符切割准确率高达97.3%。     

基于字符分割和LeNet-5网络的字符验证码识别

为了解决传统验证码识别方法效率低,精度差的问题,设计了一种先分割后识别的验证码处理方案。该方案在预处理阶段用中值滤波去噪,再利用霍夫变换对图像字符进行矫正;在字符分割阶段,利用垂直投影算法确定验证码字符块个数,以及字符坐标点,再用颜色填充算法对验证码进行初步分割,根据分割后的字符块数量对粘连字符进行二次分割;在识别阶段,我们对LeNet-5网络进行了改进,修改了输入层,并用全连接层替换了LeNet-5网络中的C5层,以此来对验证码字符进行识别;实验表明,对于非粘连验证码和粘连验证码,单张图片分割时间为0.14和0.15ms,分割准确率为98.75%和97.25%,识别准确率为99.99%和97.7%;结果表明,该算法对验证码分割和识别都有着很好的效果。     

基于步态触觉信息的图书馆智能机器人异常状态检测系统

针对当前图书馆智能机器人步态识别准确率低,导致异常状态检测效果差的问题,提出基于步态触觉信息的图书馆智能机器人异常状态检测和分类。采用基于局部空间信息加权的K-means算法对静态步态图像进行分割处理,分别构建基于改进K-means的CNN网络模型和基于时域注意力的3D残差网络模型,通过这两个模型对静态、动态步态进行特征提取和识别。实验结果表明,对比于SVM分类器,改进K-means算法的CNN网络模型静态步态识别准确率高达98.7%;3D-CNN模型的动态步态分类准确率为99.72%,均高于其他分类模型。最后结合两种算法进行异常状态检测发现,本算法的分类准确率、敏感度和特异性分别为95.42%、95.53%、94.37%。综合分析可知,提出的算法能够实现静态动态的准确识别和异常状态检测,具有一定有效性。     

基于组合特征的Bp神经网络数字识别方法

提出一种组合特征作为Bp神经网络输入层向量实现数字字符识别算法.该算法首先引入了数字字符结构特征中图段特征,并结合数字字符的行列统计特征组合成为新的特征向量;然后根据新的组合特征向量设计Bp神经网络分类器;最后对已有的数字图像样本空间中的训练样本库按照Bp神经网络分类器训练方法进行训练,并对测试样本库中的样本进行识别.根据测试实验,数字字符的识别准确率可达到94%以上.     

基于马尔科夫随机场的粘连字符串切分算法

粘连字符串模式复杂,难以通过基于传统图像处理的方法进行准确分割,针对该问题,提出一种基于机器学习的粘连字符串切分方法.包括训练和分割2个部分,对字符串之间的分割位置进行学习,对于输入的粘连字符串,利用马尔科夫随机场网络得到各点可作为分割点的概率,在概率图上使用图像分割的算法确定分割位置.实验结果表明,该算法对模拟的粘连字符串、重叠字符串和真实的手写字符串都可以得到较好的分割结果.     

基于卷积神经网络的铁轨路牌识别方法

轨道交通在我国综合交通运输体系中占有重要的地位,随着人工智能的发展,智能感知轨道交通周围环境的信息也变得越来越引人注目.本文结合深度学习与图像处理的方法,设计并实现了一种基于卷积神经网络的高铁轨道周边路牌数字识别的智能系统,该系统通过在高铁驾驶室内安装摄像头的方式采集运行前方的视频,并通过目标识别、语义分割等深度学习算法自动定位并识别路牌内的数字,从而解决了之前人工处理的繁琐和低效率.本算法整体系统由三个子模块构成,分别为目标检测模块、语义分割模块以及数字识别模块,其中目标检测模块基于SSD(Single shot MultiBox dector)模型,并对其进行了改进,使其更适用于轨道交通中的小目标识别;语义分割模块使用了全卷积的方式,对目标检测的结果进一步处理,准确得到路牌中的数字区域;数字识别模块的设计参考了著名的识别MNIST数据集的手写体识别系统,并针对路牌中数字的特点做了相应的改进,实现了对每个数字的准确识别.实验结果表明,本系统可适应白天、夜间情况下轨道交通的路况,识别的综合准确率为80.45%,其中,白天的平均识别准确率为87.98%,夜间的平均识别准确率为72.92%.     

智能车票识别系统研究

智能车票识别系统在国内尚无先例,车票文本信息的识别是此研究的关键.针对不同的文本信息给出了合理的解决方案,采用Niblack二值化算法、垂直差分投影算法、基于识别的分割等方法解决了粘连断裂字符串分割这一难题,在分割的基础上用BP神经网络进行识别;提出了一种新的基于先验知识的条码二值化算法,对128码的标准译码算法进行了改进并联合使用相关译码算法;最后综合利用数字识别结果、数据库技术、图像匹配技术解决了车票的汉字信息(站点名)识别.实验表明,提出的方法有效地完成了车票文本信息的识别,识别率较高.     

可扩展的花卉种类识别

目的 基于模式识别的花卉种类识别方法在使用不同特征或分类器时识别准确率有较大差别。本文的研究目的在于实现花卉种类识别方法的快速构建及性能评估,减轻研究人员的编程工作量,提高效率。方法 根据使用模式识别技术进行花卉种类识别的一般步骤,应用插件技术将算法中的预处理、特征提取、分类器训练、分类器识别等步骤表示成不同种类的处理器,建立可扩展的系统平台,研究人员可以通过修改各步骤所使用的处理器来修改图像处理和识别算法,并在此基础上采用流式链接方法构建算法。结果 基于本文所提出的构建方法进行算法优化,并将其用于68种花卉的识别,准确率Top1为 91.26%,Top5为98.41%。结论 流式链接方法能够对识别方法进行快速装配,有利于快速评估不同特征和分类器在花卉种类识别中的性能,适于算法的研究和优化。本文所提出的基于工作流式链接方法以及插件技术的构建方法具有灵活易用的特点,所构建的算法具有良好的可扩展性。该方法还可以推广应用于其他基于数字图像的模式识别算法研究。     

A genetic framework using contextual knowledge for segmentation and recognition of handwritten numeral strings

For the first time, a genetic framework using contextual knowledge is proposed for segmentation and recognition of unconstrained handwritten numeral strings. New algorithms have been developed to locate feature points on the string image, and to generate possible segmentation hypotheses. A genetic representation scheme is utilized to show the space of all segmentation hypotheses (chromosomes). For the evaluation of segmentation hypotheses, a novel evaluation scheme is introduced, in order to improve the outlier resistance of the system. Our genetic algorithm tries to search and evolve the population of segmentation hypotheses, and to find the one with the highest segmentation/recognition confidence. The NIST NSTRING SD19 and CENPARMI databases were used to evaluate the performance of our proposed method. Our experiments showed that proper use of contextual knowledge in segmentation, evaluation and search greatly improves the overall performance of the system. On average, our system was able to obtain correct recognition rates of 95.28% and 96.42% on handwritten numeral strings using neural network and support vector classifiers, respectively. These results compare favorably with the ones reported in the literature.     

Effects of classifier structures and training regimes on integrated segmentation and recognition of handwritten numeral strings

In integrated segmentation and recognition of character strings, the underlying classifier is trained to be resistant to noncharacters. We evaluate the performance of state-of-the-art pattern classifiers of this kind. First, we build a baseline numeral string recognition system with simple but effective presegmentation. The classification scores of the candidate patterns generated by presegmentation are combined to evaluate the segmentation paths and the optimal path is found using the beam search strategy. Three neural classifiers, two discriminative density models, and two support vector classifiers are evaluated. Each classifier has some variations depending on the training strategy: maximum likelihood, discriminative learning both with and without noncharacter samples. The string recognition performances are evaluated on the numeral string images of the NIST special database 19 and the zipcode images of the CEDAR CDROM-1. The results show that noncharacter training is crucial for neural classifiers and support vector classifiers, whereas, for the discriminative density models, the regularization of parameters is important. The string recognition results compare favorably to the best ones reported in the literature though we totally ignored the geometric context. The best results were obtained using a support vector classifier, but the neural classifiers and discriminative density models show better trade-off between accuracy and computational overhead.     

Class-specific feature polynomial classifier for pattern classification and its application to handwritten numeral recognition

The polynomial classifier (PC) that takes the binomial terms of reduced subspace features as inputs has shown superior performance to multilayer neural networks in pattern classification. In this paper, we propose a class-specific feature polynomial classifier (CFPC) that extracts class-specific features from class-specific subspaces, unlike the ordinary PC that uses a class-independent subspace. The CFPC can be viewed as a hybrid of ordinary PC and projection distance method. The class-specific features better separate one class from the others, and the incorporation of class-specific projection distance further improves the separability. The connecting weights of CFPC are efficiently learned class-by-class to minimize the mean square error on training samples. To justify the promise of CFPC, we have conducted experiments of handwritten digit recognition and numeral string recognition on the NIST Special Database 19 (SD19). The digit recognition task was also benchmarked on two standard databases USPS and MNIST. The results show that the performance of CFPC is superior to that of ordinary PC, and is competitive with support vector classifiers (SVCs).     

SEPARATION OF SINGLE-TOUCHING HANDWRITTEN NUMERAL STRINGS BASED ON STRUCTURAL FEATURES

In this paper, we develop a new method to separate single-touching handwritten numeral strings with two numerals using structural features. A binary image of a single-touching handwritten numeral string is preprocessed with an efficient algorithm for smoothing, linearization and detection of structural points of image contours. The touching region of a single-touching handwritten numeral string is determined based on distribution of the structural points in the handwritten numeral string. A candidate touching point is preselected based on the geometrical information of a special structural point in the touching region. In some cases, the left or right lateral numeral of a single-touching handwritten numeral string can be recognized. The recognition information can be utilized to correct the position of the candidate touching point. We have tested our method on image samples taken from the U.S. National Institute of Science and Technology (NIST) database. We used 500 sample images for training and obtained a correct separation rate of 99.1%. For 3287 test samples not used for training the correct separation rate was 97.2%.     

基于识别的手写数字串分割

手写数字串的分割与字符识别密切相关.采用基于识别的分割方法,在分割过程中引入识别机制识别分割碎片,将识别结果经过差值运算后置为每个识别对象的识别可信度,利用动态规划找到最佳分割路径.在训练分类器时,使用反例样本估计分类器参数,得到了性能良好的分类器.实验数据表明,利用正例和反例样本结合训练的分类器比只经过正例样本训练的分类器的识别率要高很多.     

Hierarchical classifiers based on neighbourhood criteria with adaptive computational cost

Classifiers based on neighbourhood concept require a high computational cost when the Reference Patterns Set is large. In this paper, we propose the use of hierarchical classifiers to reduce this computational cost, maintaining the hit rate in the recognition of handwritten digits. The hierarchical classifiers reach the hit rate of the best individual classifier. We have used NIST Database to carry out the experimentation, and we have worked with two test sets: in Test 1 (SD3, SD19) the hit rate is 99.54%, with a speed-up of 40.6, and in Test 2 (SD7), the hit rate is 97.51% with a speed-up of 15.7.     

Handwritten numeral recognition using gradient and curvature of gray scale image

In this paper, the authors study on the use of gradient and curvature of the gray scale character image to improve the accuracy of handwritten numeral recognition. Three procedures, based on curvature coefficient, bi-quadratic interpolation and gradient vector interpolation, are proposed for calculating the curvature of the equi-gray scale curves of an input image. Then two procedures to compose a feature vector of the gradient and the curvature are described. The efficiency of the feature vectors are tested by recognition experiments for the handwritten numeral database IPTP CDROM1 and NIST SD3 and SD7. The experimental results show the usefulness of the curvature feature and recognition rate of 99.49% and 98.25%, which are one of the highest rates ever reported for these databases (H. Kato et al., Technical Report of IEICE, PRU95-3, 1995, p. 17; R.A. Wilkinson et al., Technical Report NISTIR 4912, August 1992; J. Geist et al., Technical Report NISTIR 5452, June 1994), are achieved, respectively.     

Segmentation of single- or multiple-touching handwritten numeral string using background and foreground analysis

An approach of segmenting a single- or multiple-touching handwritten numeral string (two-digits) is proposed. Most algorithms for segmenting connected digits mainly focus on the analysis of foreground pixels. Some concentrated on the analysis of background pixels only and others are based on a recognizer. We combine background and foreground analysis to segment single- or multiple-touching handwritten numeral strings. Thinning of both foreground and background regions are first processed on the image of connected numeral strings and the feature points on foreground and background skeletons are extracted. Several possible segmentation paths are then constructed and useless strokes are removed. Finally, the parameters of geometric properties of each possible segmentation paths are determined and these parameters are analyzed by the mixture Gaussian probability function to decide the best segmentation path or reject it. Experimental results on NIST special database 19 (an update of NIST special database 3) and some other images collected by ourselves show that our algorithm can get a correct rate of 96 percent with rejection rate of 7.8 percent, which compares favorably with those reported in the literature.     

改进径向基函数神经网及其在手写体字符识别中的应用

提出一种基于半模型矢量量化（ＳＦＶＱ）技术的改进径向基函数神经网（ＩＲＢＦＮＮ）分类器,并且用于无约束手写体数字的识别。作者在模糊聚类和矢量量化的基础上利用半模糊的思想提出了半模糊矢量量化算法,并在其中加入了有监督的控制,从而使系统在聚类过程中可以确定比较合适的类别数并使聚类结果能更好地反映训练集的概率分布。以半模糊矢量量化作为预处理的改进ＲＢＦ网,应用了多尺度补偿等办法,能够充分利用训练样本集的     

Handwritten numeral recognition based on simplified structural classification and fuzzy memberships

Previous handwritten numeral recognition algorithms applied structural classification to extract geometric primitives that characterize each image, and then utilized artificial intelligence methods, like neural network or fuzzy memberships, to classify the images. We propose a handwritten numeral recognition methodology based on simplified structural classification, by using a much smaller set of primitive types, and fuzzy memberships. More specifically, based on three kinds of feature points, we first extract five kinds of primitive segments for each image. A fuzzy membership function is then used to estimate the likelihood of these primitives being close to the two vertical boundaries of the image. Finally, a tree-like classifier based on the extracted feature points, primitives and fuzzy memberships is applied to classify the numerals. With our system, handwritten numerals in NIST Special Database 19 are recognized with correct rate between 87.33% and 88.72%.