基于特征点的汉字字体识别研究

该文提出了整体分析法和个体分析法的概念,并在分析它们各自适用范围的基础上,指出个体分析法更适合于解决印刷体汉字字体识别。在此基础上,提出一种基于特征点的个体分析法来解决汉字字体识别问题,与以往方法相比,该方法具有3个优点:识别可信度可控;处理速度快;适用于多语混排情况。实验结果表明,该方法有效解决了印刷体汉字字体识别问题,其性能大大优于以往方法。     

车牌识别中字符识别的研究

对智能交通系统进行了简要的概述,分析了字符识别在车牌识别中的重要性及选取特征的标准,针对车牌中汉字与大写字母和数字采取不同的特征提取方法,最后进行实验分析并对其识别效果进行比较,发现用改进的13点特征法对汉字进行特征提取,其识别效果要好于直接用13点特征法对汉字进行识别。     

脱机手写体汉字识别综述

脱机手写体汉字识别是模式识别领域最具挑战性的课题之一。本文分析了近年来脱机手写体汉字识别的最新进展,讨论了脱机手写体汉字分割、特征提取和分类器设计等关键技术的各种主流方法,介绍了3种典型的汉字识别数据库,并提出了脱机手写体汉字识别的难点问题和今后发展的趋势,为该领域的研究者指明研究方向,共同促进脱机手写体汉字识别技术的发展。     

一种用于汉字识别的字形编码系统设计与实现

汉字字形变化多种多样,印刷体字符具有字体差异,手写字体更是没有统一的规则,识别难度非常大,当前用于汉字识别的字形编码方法大多依据字符,无法区分笔画相近的汉字。为此设计一种新的用于汉字识别的字形编码系统,介绍了汉字字形编码的原理,并给出了字形设计方案,依据编码原则,按照汉字被拆分的部件个数对汉字字形编码方案进行设计。详细介绍了汉字输入编辑器IME的结构,通过IME实现汉字的输入。依据汉字的使用频率与分布特性,通过数理统计工具设计含有汉字活动字库的操作系统,主要包括CC-DOS和MPC-DOS操作系统。实验结果表明,采用所设计系统对汉字进行识别精度较高且编码时间少、能耗低。     

基于矩和Gabor变换的手写体汉字识别方法

特征提取是手写体汉字识别的一个研究难点。本文提出了一种基于特征融合的新提取方法,即将Gabor变换和正交矩变换结合起来,用正交Zernike矩提取全局特征,用Gabor变换提取局部特征,然后使用主成分分析的方法进行特征压缩。由此得到的特征向量能从整体上和局部上反映汉字的特征。实验证明该方法是行之有效的。     

一个基于SDM改进模型的手写体汉字联想记忆系统的实现

文章针对大样本休的手与体汉字数据，指出了改进ＳＤＭ模型的必要性，提出了两点改进措施，并在此基础上采用面向对象的程序设计方法，构造了一个手写体汉字联想记忆系统。实验表明：该系统具有较好的联想记忆性能。     

识别联机手写体汉字的多级分类方法

本文提出了一种基于笔划的联机手写体汉字多级识别方法，包括拐折码笔划识别方法和笔划位置的模糊匹配方法，拐折码法具有笔划识别率高，对笔划变形的容忍度大，计算简单等特点，笔划位置匹配法不要求笔顺，对书写者的限制比较少，本方法对国标二级汉字的识别率为95％，不要求笔顺，识别速度达到实时要求。     

汉语同音字的快速识别

分析了GB2 312 - 80国标码汉字字符集的拼音分布规律。介绍了作者独创的同音汉字快速查找识别方法。     

手写汉字的四角识别方案

手写汉字的本质特征的抽取是提高识别率的关键，而识别规则的生成又依赖于大量样本的归纳学习，手写汉字的四角方案就是针对这两个问题设计并实现的。实验结果证明了该方法的可行性和优越性。     

一种基于运动图像的在线手写汉字识别方法

在分析传统联机与脱机手写汉字识别系统基础上,本文提出了一种基于运动图像的在线手写汉字识别新方法.该方法利用普通摄像头,实时地采集手写汉字墨迹的运动图像,通过图像处理技术提取手写汉字,并进行在线识别.文中给出了笔尖匹配的快速算法和基于数学形态学的汉字提取方法.最后,通过实验验证了本文方法的有效性.     

面向电磁大数据的未知雷达辐射源智能识别

基于人工智能的方法在雷达辐射源识别任务中已取得很好的效果。但随着电子信息技术的发展,将会出现越来越多的未知辐射源,其特征分布与类别都是未知的,在缺少先验知识的情况下,难以对人工智能模型进行充分的训练,使得现有的大多数方法都无法很好地完成对未知雷达辐射源的识别。为了解决上述问题,提出并阐述了可用于未知雷达辐射源识别的电磁大数据的解决方案,重点研究了基于Flink的未知雷达辐射源快速比对检索识别算法。通过对比实验证明了该方法的有效性,其识别准确率可达87.2%,当并行度设置为6时,整个互信息与K最近邻(MI-KNN)并行化算法耗时仅为4.7 s。     

3D人脸识别研究探索

2D人脸识别技术虽已成熟,但由于单一的2D图像不能提供识别所需的完整信息,故其识别精度很难进一步提高。在人脸识别过程中,特征提取是影响识别效果的一个重要环节,在分析了传统的主成分分析法和由此改进的2D PCA方法的基础上,提出了3D人脸识别方法。该方法将人脸图像分为几个部分分别进行特征提取,同时充分考虑每个部分所包含的特征信息量的多少,并在分类时赋予它们不同的权值。因此,将人脸用立体图像来表示并进行识别是目前提高人脸识别精度的前沿课题。     

Identification of humans using gait

We propose a view-based approach to recognize humans from their gait. Two different image features have been considered: the width of the outer contour of the binarized silhouette of the walking person and the entire binary silhouette itself. To obtain the observation vector from the image features, we employ two different methods. In the first method, referred to as the indirect approach, the high-dimensional image feature is transformed to a lower dimensional space by generating what we call the frame to exemplar (FED) distance. The FED vector captures both structural and dynamic traits of each individual. For compact and effective gait representation and recognition, the gait information in the FED vector sequences is captured in a hidden Markov model (HMM). In the second method, referred to as the direct approach, we work with the feature vector directly (as opposed to computing the FED) and train an HMM. We estimate the HMM parameters (specifically the observation probability B) based on the distance between the exemplars and the image features. In this way, we avoid learning high-dimensional probability density functions. The statistical nature of the HMM lends overall robustness to representation and recognition. The performance of the methods is illustrated using several databases.     

电子电路图的自动输入和识别

在这篇文章中，我们将介绍一种新型电子电路图纸自动读入系统的构成与实现。这是一个已被我们研制成功的实验系统。下面将扼要介绍其中的模式分离与图象识别技术。对于前者，我们着重介绍基于Ｆｒｅｅｍａｎ链码边缘跟踪技术的字符分离方法；而后者则包括两方面的内容：实现字符识别的人工神经网络方法与实现符号识别的决策树方法，最后给出了经识别后的输出图形。     

联合多尺度分块和协作表示的虹膜识别算法

基于稀疏表示的虹膜识别方法(SRIR)相对于传统的虹膜识别方法,在处理噪声干扰等问题,识别效果相对较好,具有较好的鲁棒性。但在样本集不足的情况下,识别性能受到影响,存在运行耗时过多、计算复杂较高的问题。针对上述问题,提出了一种联合多尺度分块和协作表示的虹膜识别算法。通过将虹膜图像按照多个尺度大小分别进行均匀分块,从而达到有效地利用虹膜特征,然后分别对每个尺度下的虹膜图像子块进行基于协作表示的识别,以降低算法耗时,最后将识别结果通过贝叶斯融合方法得到最终的分类。实验结果表明,该算法对于虹膜样本集较少的问题,比原有的SRIR方法耗时低,识别率高,复杂度低。      

Multiple instance deep learning for weakly-supervised visual object tracking

Intelligently tracking objects with varied shapes, color, lighting conditions, and backgrounds is an extremely useful application in many HCI applications, such as human body motion capture, hand gesture recognition, and virtual reality (VR) games. However, accurately tracking different objects under uncontrolled environments is a tough challenge due to the possibly dynamic object parts, varied lighting conditions, and sophisticated backgrounds. In this work, we propose a novel semantically-aware object tracking framework, wherein the key is weakly-supervised learning paradigm that optimally transfers the video-level semantic tags into various regions. More specifically, give a set of training video clips, each of which is associated with multiple video-level semantic tags, we first propose a weakly-supervised learning algorithm to transfer the semantic tags into various video regions. The key is a MIL (Zhong et al., 2020) [1]-based manifold embedding algorithm that maps the entire video regions into a semantic space, wherein the video-level semantic tags are well encoded. Afterward, for each video region, we use the semantic feature combined with the appearance feature as its representation. We designed a multi-view learning algorithm to optimally fuse the above two types of features. Based on the fused feature, we learn a probabilistic Gaussian mixture model to predict the target probability of each candidate window, where the window with the maximal probability is output as the tracking result. Comprehensive comparative results on a challenging pedestrian tracking task as well as the human hand gesture recognition have demonstrated the effectiveness of our method. Moreover, visualized tracking results have shown that non-rigid objects with moderate occlusions can be well localized by our method.     

Appearance-based periocular features in the context of face and non-ideal iris recognition

Developing newer approaches to deal with non-ideal scenarios in face and iris biometrics has been a key focus of research in recent years. The same reason motivates the study of the periocular biometrics as its use has a potential of significantly impacting the iris- and face-based recognition. In this paper, we explore the utility of the various appearance features extracted from the periocular region from different perspectives: (i) as an independent biometric modality for human identification, (ii) as a tool that can aid iris recognition in non-ideal situations in the near infra-red (NIR) spectrum, and (iii) as a possible partial face recognition technique in the visible spectrum. We employ a local appearance-based feature representation, where the periocular image is divided into spatially salient patches, appearance features are computed for each patch locally, and the local features are combined to describe the entire image. The images are matched by computing the distance between the corresponding feature representations using various distance metrics. The evaluation of the periocular region-based recognition and comparison to face recognition is performed in the visible spectrum using the FRGC face dataset. For fusion of the periocular and iris modality, we use the MBGC NIR face videos. We demonstrate that in certain non-ideal conditions encountered in our experiments, the periocular biometrics is superior to iris in the NIR spectrum. Furthermore, we also demonstrate that recognition performance of the periocular region images is comparable to that of face in the visible spectrum.     

Construction of power system transient stability equivalents using the Lyapunov function

Owing to increasing size and complexity in power systems, the study of the stability equivalent is receiving a great deal of attention. This paper describes a systematic method for the recognition of coherent machines by means of the Lyapunov function which is used for transient stability analysis. A group of generators, whose partial Lyapunov function has a small value compared with that for the whole system is aggregated into one equivalent. generator. This method does not require a long time simulation of the entire system. The parameter of the simplified system are also determined by using the Lyapunov function. The method in applied to 10-machine and 50-machine sample systems and the results are shown.     

一种RNN-T与BERT相结合的端到端语音识别模型

端到端语音识别模型由于结构简单且容易训练,已成为目前最流行的语音识别模型。然而端到端语音识别模型通常需要大量的语音-文本对进行训练,才能取得较好的识别性能。而在实际应用中收集大量配对数据既费力又昂贵,因此其无法在实际应用中被广泛使用。本文提出一种将RNN-T(Recurrent Neural Network Transducer,RNN-T)模型与BERT(Bidirectional Encoder Representations from Transformers,BERT)模型进行结合的方法来解决上述问题,其通过用BERT模型替换RNN-T中的预测网络部分,并对整个网络进行微调,从而使RNN-T模型能有效利用BERT模型中的语言学知识,进而提高模型的识别性能。在中文普通话数据集AISHELL-1上的实验结果表明,采用所提出的方法训练后的模型与基线模型相比能获得更好的识别结果。