People Recognition by RGB and NIR Analysis from Digital Image Database Using Cross-Correlation and Wavelets

This document presents a framework for recognizing people by palm vein distribution analysis using cross-correlation based signatures to obtain descriptors. Haar wavelets are useful in reducing the number of features while maintaining high recognition rates. This experiment achieved 97.5% of individuals classified correctly with two levels of Haar wavelets. This study used twelve-version of RGB and NIR (near infrared) wavelength images per individual. One hundred people were studied;therefore 4,800 instances compose the complete database. A Multilayer Perceptron (MLP) was trained to improve the recognition rate in a k-fold cross-validation test with k = 10. Classification results using MLP neural network were obtained using Weka (open source machine learning software).

SPATIAL REGULARIZATION OF CANONICAL CORRELATION ANALYSIS FOR LOW-RESOLUTION FACE RECOGNITION

Canonical correlation analysis ( CCA ) based methods for low-resolution ( LR ) face recognition involve face images with different resolutions ( or multi-resolutions ), i.e.LR and high-resolution ( HR ) .For single-resolution face recognition , researchers have shown that utilizing spatial information is beneficial to improving the recognition accuracy , mainly because the pixels of each face are not independent but spatially correlated.However , for a multi-resolution scenario , there are no related works.Therefore , a method named spatial regularization of canonical correlation analysis ( SRCCA ) is developed for LR face recognition to improve the performance of CCA by the regularization utilizing spatial information of different resolution faces.Furthermore , the impact of LR and HR spatial regularization terms on LR face recognition is analyzed through experiments.

Most relevant weighted filtering with one-step singular correlation recognition

Based on the recognition of one-step singular correlation and the remedying methods obtained before,the correlation properties of the neighborhood pixels and the characteristics of image de-noising were analyzed.A kind of most relevant weighted filtering method based on one-step singular correlation recognition(OSSC-MRWF)was put forward.The simulation experiments were done and the comparison with some commonly used methods under salt-and-pepper noises was made.The results show that the proposed method can not only effectively recognize salt-and-pepper noises and mend up the noise points,but also protect the original information such as the edge details very well.The accuracy and performance indicators are further improved considerably.

A blind modulation recognition algorithm based on cyclic spectral correlation

Cyclic spectral correlation above the bifrequency plane for the received signal was calculated by the strip spectral correlation algorithm （SSCA）and then was normalized. The result was expressed by matrix. The sum of error-square was computed between corresponding elements for the theoretical sampling matrix of all kinds of modulated signals and calculated matrix. The modulation type was recognized by exploiting the minimum value of the sum of error-square. No extracted characteristic parameter and prior information are needed for identifying the modulation type compared to the conventional methods. In addition, the new method extends the recognition scope and has high recognition probability at low SNR. The simulation results obtained by means of Monter-Carlo method proved the presented algorithm.

Facial expression recognition based on fuzzy-LDA/CCA

A novel fuzzy linear discriminant analysis method by the canonical correlation analysis （fuzzy-LDA/CCA）is presented and applied to the facial expression recognition. The fuzzy method is used to evaluate the degree of the class membership to which each training sample belongs. CCA is then used to establish the relationship between each facial image and the corresponding class membership vector, and the class membership vector of a test image is estimated using this relationship. Moreover, the fuzzy-LDA/CCA method is also generalized to deal with nonlinear discriminant analysis problems via kernel method. The performance of the proposed method is demonstrated using real data.

Intent Pattern Recognition of Lower-limb Motion Based on Mechanical Sensors

Based on the regularity nature of lower-limb motion,an intent pattern recognition approach for above-knee prosthesis is proposed in this paper. To remedy the defects of recognizer based on electromyogram(EMG), we develop a pure mechanical sensor architecture for intent pattern recognition of lower-limb motion. The sensor system is composed of an accelerometer, a gyroscope mounted on the prosthetic socket, and two pressure sensors mounted under the sole. To compensate the delay in the control of prosthesis, the signals in the stance phase are used to predict the terrain and speed in the swing phase. Specifically, the intent pattern recognizer utilizes intraclass correlation coefficient(ICC) according to the Cartesian product of walking speed and terrain. Moreover, the sensor data are fused via DempsterShafer's theory. And hidden Markov model(HMM) is used to recognize the realtime motion state with the reference of the prior step. The proposed method can infer the prosthesis user's intent of walking on different terrain, which includes level ground,stair ascent, stair descent, up and down ramp. The experiments demonstrate that the intent pattern recognizer is capable of identifying five typical terrain-modes with the rate of 95.8%. The outcome of this investigation is expected to substantially improve the control performance of powered above-knee prosthesis.

Spin-image surface matching based target recognition in laser radar range imagery

We explore the problem of in-plane rotation-invariance existing in the vertical detection of laser radar （Ladar） using the algorithm of spin-image surface matching. The method used to recognize the target in the range imagery of Ladar is time-consuming, owing to its complicated procedure, which violates the requirement of real-time target recognition in practical applications. To simplify the troublesome procedures, we improve the spin-image algorithm by introducing a statistical correlated coeff^cient into target recognition in range imagery of Ladar. The system performance is demonstrated on sixteen simulated noise range images with targets rotated through an arbitrary angle in plane. A high efficiency and an acceptable recognition rate obtained herein testify the validity of the improved algorithm for practical applications. The proposed algorithm not only solves the problem of in-plane rotation-invariance rationally, but also meets the real-time requirement. This paper ends with a comparison of the proposed method and the previous one.

TEMPLATE MATCHING ALGORITHM OF RADAR BEAM SCAN TYPE RECOGNITION

Phased array radar has been applied broadly because of its sound performance.But signal of phased array radar is of a wide variety of types.Therefore,recognition of phased array radar is the most puzzling aspect of the whole emitter identification domain.To solve the problem,the article proposes the method that identifies phased array radar by pulse amplitude information,and studies the phased array radar,models transmit signal of them,and receiving signal by radar countermeasure reconnaissance receiver.From constructing template of pulse train's amplitude vector of mechanical scanning radar,computing distance of samples and standard template,finding threshold of the template matching arithmetic,the article puts forward the template matching algorithm of radar beam scan type recognition to identify phased array radar automatically.

LOCAL CORRELATION DISCRIMINANT ANALYSIS AND ITS SEMI-SUPERVISED EXTENSION

Considering limitations of Linear Discriminant Analysis (LDA) and Marginal Fisher Analysis (MFA), a novel discriminant analysis called Local Correlation Discriminant Analysis (LCDA) is proposed in this paper. The main idea behind LCDA is to use more robust similarity measure, correlation metric, to measure the local similarity between image data. This results in better classifi-cation performance. In addition, to further improve the discriminant power of LCDA, we extend LCDA to semi-supervised case, which can make use of both labeled and unlabeled data to perform dis-criminant analysis. Extensive experimental results on ORL and AR face databases demonstrate that the proposed LCDA and its semi-supervised version are superior to Principal Component Analysis (PCA), LDA, CEA, and MFA.

Liver fibrosis recognition using multi-compression elastography technique

Liver fibrosis recognition is an important issue in diagnostic imaging. The accurate estimation of liver fibrosis stages is important to establish prognosis and to guide appropriate treatment decisions. Liver biopsy has been for many years the reference procedure to assess histological definition for liver diseases. But biopsy measurement is an invasive method besides it takes large time. So, fast and improved methods are needed. Using elastography technology, a correlation technique can be used to calculate the displacement of liver tissue after it has suffered a compression force. This displacement is related to tissue stiffness, and liver fibrosis can be classified into stages according to that displacement. The value of compression force affects the displacement of tissue and so affects the results of the liver fibrosis diagnosing. By using finite element method, liver fibrosis can be recognized directly within a short time. The proposed work succeeded in recognizing liver fibrosis by a percent reached in average to 86.67% on a simulation environment.

Biometric feature extraction using local fractal auto-correlation

Image texture feature extraction is a classical means for biometric recognition. To extract effective texture feature for matching, we utilize local fractal auto-correlation to construct an effective image texture descriptor. Three main steps are involved in the proposed scheme： （i） using two-dimensional Gabor filter to extract the texture features of biometric images; （ii） calculating the local fractal dimension of Gabor feature under different orientations and scales using fractal auto-correlation algorithm; and （iii） linking the local fractal dimension of Gabor feature under different orientations and scales into a big vector for matching. Experiments and analyses show our proposed scheme is an efficient biometric feature extraction approach.

Label correlation for partial label learning

Partial label learning aims to learn a multi-class classifier,where each training example corresponds to a set of candidate labels among which only one is correct.Most studies in the label space have only focused on the difference between candidate labels and non-candidate labels.So far,however,there has been little discussion about the label correlation in the partial label learning.This paper begins with a research on the label correlation,followed by the establishment of a unified framework that integrates the label correlation,the adaptive graph,and the semantic difference maximization criterion.This work generates fresh insight into the acquisition of the learning information from the label space.Specifically,the label correlation is calculated from the candidate label set and is utilized to obtain the similarity of each pair of instances in the label space.After that,the labeling confidence for each instance is updated by the smoothness assumption that two instances should be similar outputs in the label space if they are close in the feature space.At last,an effective optimization program is utilized to solve the unified framework.Extensive experiments on artificial and real-world data sets indicate the superiority of our proposed method to state-of-art partial label learning methods.

Iris Feature Extraction and Recognition Based on Fractal Dimension of Grayscale Extremums

This paper presented an individual recognition algorithm for human iris using fractal dimension of grayscale extremums for feature extraction.Firstly,iris region was localized from an eye image with modified circle detector stemmed from Daugman’s integro-differential operator.Then,segmentation was used to extract the iris and to exclude occlusion from eyelids and eyelashes.The extracted iris was normalized and mapped to polar coordinates for matching.In feature encoding,a new approach based on fractal dimension of grayscale extremums was designed to extract textural features of iris.Finally,a normalized correlation classifier was employed to determine the agreement of two iris feature templates,and the feature template was rotated left and right to avoid the interference from rotation of eyes and tilting of head.The experimental results show that fractal dimension of grayscale extremums can extract textural features from iris image effectively,and the proposed recognition algorithm is accurate and efficient.The proposed algorithm was tested on CASIA-IrisV3-Interval iris database and the performance was evaluated based on the analysis of both False Accept Rate(FAR)and False Reject Rate(FRR)curves.Experimental results show that the proposed iris recognition algorithm is effective and efficient.

Correlation Analysis of Gridding of Earthquakes in Datong and Its Surrounding Areas and Implication for Earthquake Prediction

On the basis of the earthquake （gL I〉3.0） catalog in North China from 1970 to 2009, the pattern of temporal and spatial distribution of medium-small earthquakes in Datong and its surrounding areas is studied by correlation analysis with a focus on its anomaly before moderate and strong earthquakes. With different spatial scales, temporal scales and time steps, the spatial distribution of earthquakes is converted to a sequence, then the correlation coefficients between the spatial distribution of medium-small earthquakes in a long-term and a longer time are calculated for the analysis of anomalies before moderate and strong earthquakes. In the study region center on the epicenter of the 1989 Datong- Yanggao earthquake （Ms5. 9） within a radius of less than 0.8~, with the time length of 3600 days, the longer time length of 3700 days, and the time step of 100 days, the correlation coefficient from 1980 to 2009 is steady between 0.94 and 1.00, but there were anomalies with values less than 0. 94 in the 2 years before the 1989 Datong-Yanggao earthquake （Ms 5.9）, the 1991 Datong earthquake （ Ms 5.8） and 1999 Hunyuan earthquake （Ms 5. 6 ）, which indicates the spatial distribution of a medium-small earthquake is very different from steady background seismicity. The implication for earthquake prediction from the anomaly of the correlation coefficient is also discussed with the three conclusions： （1） Before moderate and strong earthquakes in Datong and its surrounding areas, the obvious change of spatial distribution patterns of medium-small earthquake can be a kind of seismic precursor of the 2-year time scale for the prediction of an earthquake＇s time. （2） As the study region is restricted within a radius of less than 0. 8~, the result of correlation analysis is also good for the prediction of an earthquake＇s location. （3） The method of correlation analysis in this paper helps recognize the anomaly of spatial distribution of medium-small earthquake.

地磁异常识别技术在松原5.7级地震前的应用研究

运用地磁异常识别技术对2018年5月28日松原5.7级地震前的地磁异常展开研究。采用加卸载响应比法、逐日比法、日变化空间相关法、异常叠加法对吉林省三岗台、通化台及东北地区邻近的大连台、营口台、朝阳台、铁岭台等台站,运用松原5.7级地震前18个月内的地磁预处理分钟值数据进行计算分析和异常判定识别。研究结果表明:在加卸载响应比法、逐日比法和日变化空间相关法的基础上,运用异常叠加法提取的地磁异常,其预测的位置与真实值更加接近,预测结果对省内地震具有重要的指示意义。

基于特征融合的无人船目标识别系统设计

通过特征融合可获取全面的目标特征信息,利于提升目标识别的稳定性,为此设计基于特征融合的无人船目标识别系统。利用无人船搭载红外热成像仪与可见光摄像头,采集目标红外与可见光图像;通过处理器和可编程逻辑控制器,设计特征提取模块,用于提取红外与可见光图像的无人船目标特征;特征融合模块利用典型相关分析理论,融合红外与可见光图像的无人船目标特征;目标识别模块通过径向基函数网络,结合特征融合结果,输出无人船目标识别结果。实验结果证明,该系统可有效采集无人船目标的红外与可见光图像,完成特征提取;该系统具备较优的特征融合效果,并精准实现无人船目标识别。

基于PSD特征的FBCCA脑电信号识别方法

当前基于稳态视觉诱发电位(steady-state visual evoked potential,SSVEP)的脑机接口(brain-computer interfaces,BCIs)使用的都是单一识别算法,针对不同时间长度的识别准确率较低。提出了一种基于滤波器组的典型相关分析(filter bank canonical correlation analysis,FBCCA)与功率谱密度(power spectral density,PSD)分析相结合的SSVEP识别算法,可以提高SSVEP识别的普适性与准确率。该方法使用FBCCA寻找高相似度的参考频率信号,再通过多组PSD分析来锁定最终的响应频率,完成频率识别。该方法无需经过训练就能得到较高的识别准确率。实验结果表明:在刺激时长为1 s时,该方法能达到86.61%的准确率,比PSD分析方法提升了5.44%,比典型相关性分析方法(canonical correlation analysis,CCA)提升了10.38%的准确率,比FBCCA提升了8.86%的准确率。

时空多尺度关联特征融合的二维卷积网络细粒度动作识别模型

针对传统二维(2D)卷积网络提取时空特征尺度单一以及对细粒度动作数据集中帧与帧之间的远程时间关联信息利用不足的问题,本文提出时空多尺度关联特征融合的2D卷积网络细粒度动作识别模型。首先,为建模视频多尺度空间关联以加强对细粒度视频数据的空间表征能力,模型使用多尺度“特征压缩、特征激发”方式,使网络所提取空间特征更加丰富有效。然后,为充分利用细粒度视频数据时间维度上的运动信息,本文引入时间窗口自注意力机制,利用自注意力机制强大的远程依赖建模能力同时只在时间维度上进行自注意力操作,以较低计算成本建模远程时间依赖关系。最后,考虑到所提取时空特征对不同类型动作分类的贡献不均等,本文引入自适应特征融合模块,为特征动态赋予不同权重实现自适应特征融合。模型在2个细粒度动作识别数据集Diving48和Something-somethingV1上识别准确率分别达到86.0%和46.9%,分别使原始主干网络识别准确率提升3.8%和1.3%。实验结果表明,在只使用视频帧信息作为输入的情况下,本模型达到与现有基于Transformer和三维卷积神经网络(3D CNN)算法相当的识别准确率。

基于CCA融合FFT的SSVEP脑机接口分类算法

为解决多目标刺激范式的稳态视觉诱发电位脑电信号识别准确率低和信息传输率低的问题,提出了一种快速傅里叶变换同典型相关分析相结合的方法,通过快速傅里叶变换将信号训练成对应频率的训练模板,并作为参考信号与实时采集的信号进行典型相关分析来计算频率的识别准确率。6名受试者参与并完成了180组实验,在时间窗口长度为1.5 s的条件下,基于快速傅里叶变换-典型相关分析的稳态视觉诱发电位信号识别算法的平均识别准确率为93.98%,比典型相关分析算法提升了14.75%,信息传输率为62.30 bit·min^(-1),比典型相关分析算法提升了55.63%。实验结果表明,快速傅里叶变换-典型相关分析算法性能更优。

雷达有源干扰的多域特征参数关联智能识别算法

针对复杂干扰场景下参数变化范围大、决策树特征适应性较差等问题,本文提出了一种基于多域特征参数关联的雷达有源干扰智能识别方法。首先,针对干扰信号在时域、频域、脉压等多个域中的特点,设计了相应的特征参数。这些特征参数能够全面地描述干扰信号的特征。接着,采用随机森林的平均信息增益来选择不同干扰参数条件下的特征参数,并将它们进行关联。然后,结合多域关联特征和典型的神经网络,提出了一种改进的ResNet18网络。通过利用多域关联特征,该网络能够更准确地学习干扰信号的特征,并进行智能识别。最后设置大范围参数的干扰样本训练改进的ResNet18网络,优化其泛化性能。通过计算机仿真实验,本文的方法在大参数范围的有源干扰识别正确率达到了98%以上。而且经过改进后的ResNet18网络参数总量仅为原网络的1/17,大大减少了网络的复杂度。综上所述,本文提出的基于多域特征参数关联的雷达有源干扰智能识别方法具有较高的识别准确性和较低的计算复杂度,可以有效应用于复杂干扰场景中。