General logarithm difference model for severe illumination variation face recognition

Multimedia Tools and Applications - Logarithm difference is the subtraction of the center pixel and its neighbor in the face local region. The commonly-used assumption of the illumination invariant...     

利用SVM进行车型识别

为了提高车辆图像的识别率,提出了利用支持向量机（SVM）理论进行轿车车型识别方法.SVM能够解决线性及非线性分类问题,以较少的支持向量确定分类面,对样本数量及维数不敏感.基于颜色直方图及惯性比确定的图像特征具有平移、旋转和尺度不变性,可以用来确定SVM的最优分类面,并由此识别车型.     

基于多尺度MRF结合超级耦合变换的姿态变化人脸识别

针对马尔可夫随机场模型(MRF)在姿态变化人脸识别中复杂度较高的问题,提出了一种多尺度MRF算法,通过串联超级耦合变换(SCT)优化多尺度MRF过程的子问题,明显降低了MRF的计算复杂度。首先,通过两层图像完成对称匹配;然后,利用多尺度局部二值模式(MLBP)直方图特征生成距离比;最后,比较获得最终比值,配对每对区域。在户外标记人脸数据库(LFW)、XM2VTS数据库和CMU-PIE数据库上进行实验,实验结果表明,相比现有的优秀算法,本文算法在识别率和执行速度方面均表现优越,在图形处理器上基本满足实际应用需求。     

基于三维人脸建模的多姿态人脸识别

待匹配人脸图像与原始图像存在姿态和光照的差异,是自动人脸识别的两个主要瓶颈问题.给出了采用三维人脸模型来解决人脸姿态的变化对人脸识别的影响问题.通过正侧面图像,利用B样条曲线与径向基函数相结合的方法进行三维人脸重建,生成三维人脸模型库.分别计算待匹配人脸图像的3个自由度,快速估计出人脸的姿态;结合待匹配人脸图像的姿态参数及三维人脸模型库,获得与待匹配图像相同姿态的三维人脸模型库中的二维人脸图像.最后完成了相同人脸姿态的二维人脸识别.实验结果证明,该方法无需复杂的设备、简单易行、识别时间短,是一种非常实用的解决人脸姿态问题的识别方法.     

基于Candide-3模型的姿态表情人脸识别研究

针对姿态表情严重影响人脸识别准确率的问题,基于Candide-3模型的简化,提出了形状表情关键点拟合的人脸几何结构重建和基于三角网格模型的纹理映射的方法,该方法确定关键特征点,根据人脸的几何结构信息确定姿态角,提取Candide-3模型形状表情对应点,调整模型参数,进行几何结构重建;对几何结构中每个三角网格模型进行纹理影射,得到逼真的特定人脸模型.实验结果表明,该方法提高了人脸重建速度,达到减弱姿态表情对人脸识别影响的目的.     

Car plate recognition by whole 2-D image

This study did not extract the characters in the car plate separately. Instead, a whole 2-D car plate was processed as an entity; subsequently, it was recognized. In this study a ring-to-line mapping technique is used to map a car plate image to several straight-line signals. Moreover, the vector magnitude invariant transform technique is used to transfer a car plate signal to an invariant vector magnitude for comparison, a procedure that can solve the image-rotation problem. Various vertical magnitude strips are generated to manage the image-shifting problem. The algorithm developed in this study can precisely identify car plate images.     

Car make and model recognition under limited lighting conditions at night

Car make and model recognition (CMMR) has become an important part of intelligent transport systems. Information provided by CMMR can be utilized when license plate numbers cannot be identified or fake number plates are used. CMMR can also be used when a certain model of a vehicle is required to be automatically identified by cameras. The majority of existing CMMR methods are designed to be used only in daytime when most of the car features can be easily seen. Few methods have been developed to cope with limited lighting conditions at night where many vehicle features cannot be detected. The aim of this work was to identify car make and model at night by using available rear view features. This paper presents a one-class classifier ensemble designed to identify a particular car model of interest from other models. The combination of salient geographical and shape features of taillights and license plates from the rear view is extracted and used in the recognition process. The majority vote from support vector machine, decision tree, and k-nearest neighbors is applied to verify a target model in the classification process. The experiments on 421 car makes and models captured under limited lighting conditions at night show the classification accuracy rate at about 93 %.     

Shape recognition by a scale-invariant model

In this article, we considered the recognition of unknown shapes by maximum likelihood methods. The contour of a shape is represented by its centroidal profile, and it is fitted by a circular autoregressive model. Two different shape recognition problems are considered: the decision on the similarity of two unknown shapes, and the classification of an unknown shape as one of many known shapes. Maximum likelihood decision rules for these two cases are derived. The decision rules are invariant to translation, rotation, and size change after normalizing the estimates. The developed algorithms are applied to classify eight classes of machine parts and eight classes of aircraft shapes. For each class, 60 to 80 samples are generated by rotating and dilating the original shape. In the experiment, more than 98% of machine parts are classified correctly, and more than 97% of aircraft shapes are correctly classified. This result is better than previous model-based approaches. Partially supported by the National Science Foundation under the grant IRI-8809391.     

Using error decay prediction to overcome practical issues of deep active learning for named entity recognition

Machine Learning - Existing deep active learning algorithms achieve impressive sampling efficiency on natural language processing tasks. However, they exhibit several weaknesses in practice,...     

云模型优化LSF调度算法的研究

为了解决LSF调度算法在实时调度中由颠簸现象引起的调度实时性差、浪费系统资源的问题,在LSF算法中引入一个任务重要度系数,采用云模型对任务重要度系数和裕度进行定量表示,并通过由重要度系数云和裕度云两个任务特征参数云模型共同确定的二维云模型,为每个任务设定一个抢占阈值,当某一就绪任务要抢占当前任务时,必须要满足它的优先级高于当前任务的抢占阈值.仿真结果表明,采用云模型优化后的LSF算法不仅有效解决了颠簸现象,而且能使紧急且重要的任务优先运行.     

Music symbol recognition by a LAG-based combination model

Most of optical music recognition (OMR) systems work under the assumption that the input image is scanner-based. However, we propose in this paper, camera based OMR system. Camera based OMR has a challengeable work in un-controlled environment such as a light, perspective, curved, transparency distortions and uneven staff-lines which tend to incur more frequently. In addition, the loss in performance of binarization methods, line thickness variation and space variation between lines are inevitable. In order to solve these problems, we propose a novel and effective staff-line removal method based on following three main ideas. First, a state-of-the-art staff-line detection method, Stable Path, is used to extract staff-line skeletons of the music score. Second, a line adjacency graph (LAG) model is exploited in a different manner over segmentation to cluster pixel runs generated from the run-length encoding (RLE) of an music score image. Third, a two-pass staff-line removal pipeline called filament filtering is applied to remove clusters lying on the staff-line. A music symbol is comprised of several parts so-called primitives, but the combination of these parts to form music symbol is unlimited. It causes difficulty applying the state-of-the-art method for music symbol recognition. To overcome these challenges and deal with primitive parts separately, we proposed a combination model which consists of LAG model, Graph model, and Set model as a framework for music symbol recognition. Our method shows impressive results on music score images captured from cameras, and gives high performance when applied to the ICDAR/GREC 2013 database, and a Gamera synthetic database. We have compared to some commercial software and proved the expediency and efficiency of the proposed method.     

Pattern recognition by an optical thin-film multilayer model

This paper describes a computational learning model inspired by the technology of optical thin-film multilayers from the field of optics. With the thicknesses of thin-film layers serving as adjustable “weights” for the computation, the optical thin-film multilayer (OTFM) model is capable of approximating virtually any kind of nonlinear mapping. This paper describes the architecture of the model and how it can be used as a computational learning model. Some sample simulation calculations that are typical of connectionist models, including a pattern recognition of alphabetic characters, iris plant classification, and time series modelling of a gas furnace process, are given to demonstrate the model’s learning capability. This revised version was published online in June 2006 with corrections to the Cover Date.     

KIII模型及其在人脸识别中的应用

对一种模拟嗅觉神经系统的仿生模型、KIII模型、及其在人脸识别中的应用进行了研究。论文首先分析了KIII模型模式识别原理,然后基于人脸识别对模型的模式容量问题进行了研究。使用子图特征组合算法提取人脸特征后,在ORL人脸数据库上的实验结果表明,KIII模型的模式容量至少可以达到40,这也更新了之前认为KIII模型模式容量约为26的观点。     

Automatic compilation of language resources for named entity recognition in Turkish by utilizing Wikipedia article titles

We present an automatic approach to compile language resources for named entity recognition (NER) in Turkish by utilizing Wikipedia article titles. First, a subset of the article titles is annotated with the basic named entity types. This subset is then utilized as training data to automatically classify the remaining titles by employing the k-nearest neighbor algorithm, leading to the construction of a significant lexical resource set for Turkish NER. Experiments on different text genres are conducted after extending an existing NER system with the resources and the results obtained confirm that the resources contribute to NER on different genres.     

基于Gabor小波和模型自适应的鲁棒人脸识别方法*

提出了一种基于Gabor小波人脸特征和模型自适应算法的新鲁棒人脸识别方法。该方法在真实识别前,通过用与真实识别相同的环境条件下所获得的人脸图像数据对原始模型进行更新补偿,实现了模型自适应。该模型自适应更新算法是加性的,其具有较低的时间和空间复杂度。通过模型自适应更新,新方法可以有效地减少模型和识别数据间的失配,从而提高识别率。在AT&T和MITCBCL人脸数据库上的测试结果表明,该方法是有效的。     

A novel statistical generative model dedicated to face recognition

In this paper, a novel statistical generative model to describe a face is presented, and is applied to the face authentication task. Classical generative models used so far in face recognition, such as Gaussian Mixture Models (GMMs) and Hidden Markov Models (HMMs) for instance, are making strong assumptions on the observations derived from a face image. Indeed, such models usually assume that local observations are independent, which is obviously not the case in a face. The presented model hence proposes to encode relationships between salient facial features by using a static Bayesian Network. Since robustness against imprecisely located faces is of great concern in a real-world scenario, authentication results are presented using automatically localised faces. Experiments conducted on the XM2VTS and the BANCA databases showed that the proposed approach is suitable for this task, since it reaches state-of-the-art results. We compare our model to baseline appearance-based systems (Eigenfaces and Fisherfaces) but also to classical generative models, namely GMM, HMM and pseudo-2DHMM.     

A model validation approach to texture recognition and inpainting

A new approach to texture recognition and inpainting problems is proposed. The approach is based on the robust model validation and state estimation techniques. The proposed solutions require the modeling of textures by using uncertain dynamical systems. We propose a new modeling method which is efficient in terms of computational and memory requirements. The main aspects of the modeling method include system identification and order reduction of marginally stable uncertain discrete-time systems. To demonstrate the results, both static-image textures and video textures (also known as dynamic textures) are considered.