A Highly Accurate Face Recognition System Using Filtering Correlation

The authors previously constructed a highly accurate fast face recognition optical correlator (FARCO) [E. Watanabe and K. Kodate: Opt. Rev. 12 (2005) 460], and subsequently developed an improved, super high-speed FARCO (S-FARCO), which is able to process several hundred thousand frames per second. The principal advantage of our new system is its wide applicability to any correlation scheme. Three different configurations were proposed, each depending on correlation speed. This paper describes and evaluates a software correlation filter. The face recognition function proved highly accurate, seeing that a low-resolution facial image size (64 × 64 pixels) has been successfully implemented. An operation speed of less than 10 ms was achieved using a personal computer with a central processing unit (CPU) of 3 GHz and 2 GB memory. When we applied the software correlation filter to a high-security cellular phone face recognition system, experiments on 30 female students over a period of three months yielded low error rates: 0% false acceptance rate and 2% false rejection rate. Therefore, the filtering correlation works effectively when applied to low resolution images such as web-based images or faces captured by a monitoring camera.     

Conjugate Image Plane Correlator with Holographic Disk Memory

A conjugate image plane correlator with holographic disk memory is proposed. Optical correlation between conjugate images reconstructed from a holographic disk and an input image on liquid-crystal television is executed with the rotation of the disk. Regardless of Fourier hologram recording with the pseudorandom diffuser, it is found possible to take out the diffuser from the original hologram recording scheme using an image reconstruction process and to get correlation signals between input and reconstructed conjugate images in the output plane of a two-lens imaging system. Generation of conjugate replicas with high contrast causes exact matching with an input image which results in high recognition performance for autocorrelation signals. The transfer function of an optical system can be controlled by adjustment of either hologram size or hologram area illuminated with a laser beam. Hence, the output intensity distribution can be adjusted by selecting a proper pupil function and the size of an output pupil defined by the input pupil size and the optical system magnification factor. The real-time character recognition by optical parallel high-speed processing for two dimensional images with position normalization is demonstrated.     

光学相关器在自动目标识别中的应用

为使光学相关器更好地应用于自动目标识别领域,保证在较高识别速度下,具备良好的识别效果,介绍了应用于自动目标识别领域的联合变换相关器和匹配滤波相关器.针对匹配滤波相关器实现过程中的两项关键技术,在滤波器编码技术上提出了构造等相关峰综合鉴别函数以解决畸变不变识别问题.在相关峰识别技术上,针对阈值法存在的不足,提出利用BP神经网络对相关平面相关峰进行识别,采用划分感兴趣区域的方法减少了输入层神经元数目,简化了神经网络结构.借助搭建的光学相关器系统验证了该方法可对相关信号和噪声进行有效分类,提高了光学相关器的识别效果.     

基于体全息光学相关器的身份认证复合检索方法

体全息相关器具有高速、并行和多通道等特点。基于体全息相关器大幅面、高分辨率、二维数据页结构,将身份认证图文信息进行编码及格式化,组合为适用于光学相关器的图像模板。数据库的图像模板以体全息图的形式存储在铌酸锂晶体中,识别时将待识别的全部或部分身份认证信息输入,光学相关器可快速返回输入图像与所有库图像模板的相关值,从而完成对整个数据库的快速复合检索。介绍了该方法的工作原理、理论模型、编码和组页方法以及复合检索工作流程,给出了演示实验结果。     

Optical Fractional Derivative Matched Correlator

We suggest a new optical matched correlator which combines the fractional derivative with the conventional correlation for application to space shift-invariant pattern recognition. This new correlator possesses the favorable feature of being able to enhance the discrimination capability and anti-interference to noise ability over that of the conventional matched spatial filter. Varying the fractional order enables one to trade off between the light efficiency and the discrimination ability of the correlator.     

Optimal Incoherent Filtering for Distortion-Invariant Image Recognition

An optimal incoherent filter is proposed for distortion-invariant correlation. The optical transfer function (OTF) of the correlator is specified as a realizable phase-only term which is optimized by the simulated annealing algorithm. The phase-only OTF is used in image recognition to produce a sharp peak response and can be generated by dual-filter synthesis. Image recognition is achieved by subtracting two spatially incoherent correlation outputs produced from a positive and negative filters with an input image. Computer simulations show that the proposed correlator can yield a sharp correlation peak with excellent distortion tolerance.Presented at 1996 International Topical Meeting on Optical Computing (OC ‘96), April 21–25, Sendai, Japan     

光电混合光盘存储二值化联合变换相关器

本文提出了基于光盘存储的单空间光调制器光电混合联合变换相关系统,该系统采用视频光盘建立参考图象库,充分结合了联合变换相关器的内在优点和光盘系统的高信息存储能力,能实时地给出相关识别结果,是适合于机器人视觉的一种有效系统。文末给出了计算机模拟结果和初步的光学实验结果。     

Wavelet-modified fringe-adjusted joint transform correlator

In this paper, we implement a wavelet-modified fringe-adjusted joint transform correlator (JTC) for real-time target recognition applications. In real-time situation the input scene is captured using a charge-coupled device (CCD) camera. The obtained joint power spectrum is multiplied by a pre-synthesized fringe-adjusted filter and the resultant function is processed with an appropriately scaled wavelet filter. Three performance measure parameters: correlation peak intensity, peak-to-sidelobe ratio, and signal-to-clutter ratio have been calculated for fringe-adjusted joint transform correlator (FJTC) and wavelet-modified fringe-adjusted joint transform correlator (WFJTC). The WFJTC has been found to yield better results in comparison to conventional FJTC. To suppress the undesired strong dc, the resultant function is differentiated. Differential processing wavelet-modified fringe-adjusted joint power spectrum removes the zero-order spectra and hence improves the detection efficiency. To focus the correlation terms in different planes in order to capture one of the desired autocorrelation peaks and discard the strong dc and another autocorrelation peak, chirp-encoding technique has also been applied. Targets with Gaussian and speckle noise have also been used to check the correlation outputs. Computer simulation and experimental results are presented.     

Binary phase only filters for rotation and scale invariant pattern recognition with the joint transform correlator

The joint transform correlator (JTC) is one of the two main optical image processing architecture which provides a highly effective way of comparing images in a wide range of applications. Traditionally, an optical correlator is used to compare an unknown input scene with a pre-captured reference image library, to detect if the reference occurs within the input. Strength of the correlation signal decreases rapidly as the input object rotates or varies in scale relative to the reference object. The aim of this paper is to overcome the intolerance of the JTC to rotation and scale changes in the target image. Many JTC systems are constructed with the use of ferroelectric liquid crystal (FLC) spatial light modulators (SLMs) as they provide fast two-dimensional binary modulation of coherent light. Due to the binary nature of the FLC SLMs used in the JTC systems, any image addressed to the device need to have some form of thresholding. Carefully thresholding the grey scale input plane and the joint power spectrum (JPS) has significant effect on the quality of correlation peaks and zero order (DC) noise. A new thresholding technique to binarise the JPS has been developed and implemented optically. This algorithm selectively enhances the desirable fringes in the JPS which provide correlation peaks of higher intensity. Zero order noise is further reduced when compared to existing thresholding techniques.Keeping in mind the architecture of the JTC and limitations of FLC SLMs, a new technique to design rotation and scale invariant binary phase only filters for the JTC architecture is presented. Filers design with this technique have limited dynamic range, higher discriminability among target and non-target objects, and convenience for implementation on FLC SLMs. Simulation and experiments shows excellent results of various rotation and scale invariant filters designed with this technique. A rotation invariant filter is needed for various machine vision applications of the JTC. By fixing the distance between camera and input object, the scale sensitivity of the correlator can be avoided. In contrast to the industrial machine vision applications, scale factor is very important factor for the applications of the JTC systems in defence and security. A security system using a scale invariant JTC will be able to detect a target object well in advance and will provide more time to take a decision.     

利用液晶光阀实现二值化联合变换相关器的研究

二值化联合变换相关器，通过对联合变换的功率谱进行二值化操作可以提高联合变换相关器的图像识别。本文正是基于此观点，首次提出利用液晶光阀本身的饱和区实现反转二值化联合变换功率谱。     

多目标识别的联合变换相关器的研究

提出了一种可用于多目标识别的联合变换相关器。为改善相关信号的性能,对功率谱作了优化处理。为消除相关面上的零级项和目标间的相关项,可用联合功率谱减去纯目标输入的功率谱和参考图像的功率谱;为增强和锐化相关峰,将相减的功率谱作指数函数滤波处理。分析了指数滤波参数对相关结果的影响。计算机模拟结果表明,这种相关器所输出的相关信号比经典联合变换相关器和二元联合变换相关器输出的相关信号更好,互相关得到了抑制,自相关得到了增强,具有很好的抗噪能力。     

新型平面集成光学相关器设计与分析

基于微光学与二元光学技术,设计了一种新型的平面折叠式光学相关器,将相关器中组成4f系统的傅里叶变换透镜、全息匹配滤波器、输入面和输出集成到一个平面器件上,与另一平面偏振反射镜构成一个反射型的平面光学系统.讨论了相关器的系统结构设计、系统参数的设计,给出了具体的设计模型,计算出其体积可减小至约12 cm3.利用菲涅耳衍射数值模拟方法对系统的图像识别过程进行了仿真,得到了良好的相关输出结果,证实平面集成光学相关器具有很好的识别性能.     

Engineering implementation of a hybrid optical correlator using SLMs

Optical correlators have been widely used for such applications as automatic target identification, invariant pattern recognition and machine vision. In hybrid optical correlators, Spatial light modulators (SLMs) are used to dynamically update the input and the filter. However, implementation is difficult because of the interference of “ghost” images; rigorous requirements for filter registration, accurate and effective filter sizing and complex valued filtering functions. This paper proposes some practical approaches to handle these issues and provides engineering details of how to build a hybrid optical correlator, where a SLM serves as an updateable intensity filter for real-time pattern recognition.     

实时高效联合变换相关器

提出一种利用光学波分复用器实现实时高效联合变换相关的方案。通过计算机控制把输入面分为多个大小相等的窗口，分别写入参考与目标图像，利用光学波分复用器于相关系统中，使对应不同窗口的读出光学信息在第一个焦面上形成联合功率谱阵列，在输出面上可以实时地获得增强了相关信号。     

修正振幅调制的光电混合圆谐联合变换相关器作旋转不变的目标检测

提出一种修正振幅调制的光电混合圆谐联合变换相关器作旋转不变的目标检测的实验系统。参考图像（即圆谐展开分量的实部和虚部）、圆盘状的局部偏置函数以及目标图像同时显示于输入面。此方法对联合功率谱作了修正,先将联合功率谱减去纯输入景物（含局部偏置函数）的功率谱和参考图像（含局部偏置函数）的功率谱,再加上局部偏置函数的功率谱,然后将所得修正的联合功率谱乘以振幅调制滤波函数。这种方法能产生比普通的圆谐联合变换     

旋转不变二值化联合变换相关识别方法

介绍一种基于二值化联合变换相关的综合鉴别函数算法，用等相关峰及Gram-schmidt展开获得的综合鉴别函数对迭代算法进一步优化，获得了适用于二值化联合变换相关的综合鉴别函数。采用这种算法制作的综合鉴别滤波器既有相关峰值高和相关峰尖锐等特点,又能识别旋转图像。计算机模拟结果表明：这种相关器在信噪比、光能效率、相关峰尖锐度和鉴别能力上都有一定程度地提高，实现了较高鉴别率旋转不变光学模式识别。     

Use of a joint transform correlator architecture for study of speckle displacements

The optical speckle–displacement correlation technique (OSDCT) is proposed for study of in-plane speckle displacements. The joint transform correlator architecture is used to evaluate the displacements of speckle subimages of strained surfaces. The procedure of the correlation response producing as a result of cross-correlation of strainless and strained surfaces in a conventional joint transform correlator (JTC) is considered. The convenient techniques for a joint power spectrum transformation are selected. The systematic and random errors of a mean speckle pattern displacement of a rigid body motion calculated by using the OSDCT and the digital speckle–displacement measurement technique are compared. The robustness of some JTC versions (a conventional JTC, a JTC with median thresholding, a JTC with subset median thresholding, and a fringe adjusted filter JTC) to input noise of speckle patterns and output noise of a correlation response is studied.     

实时声光图像相关器及其在模试识别中的应用

介绍一种实时声光图像相关器。在该系统中，一个声光偏转器和一个激光二极管阵列分别被用作输人图像和滤波器的电光转换器件。二维相关由工作于时间延迟积分模式下的一个二维ＣＣＤ探测器完成。最后描述这种相关器在模式识别中的应用。     

Wavelet-modified maximum average correlation height filter for out-of-plane rotation invariance

In this paper, we implement a wavelet-modified maximum average correlation height (WaveMACH) filter for recognition of 0-360° out-of-plane rotated targets employing hybrid digital-optical correlator architecture. A chirp function is used with the synthesized WaveMACH filter, which focuses the correlation signals in three different planes, thereby providing the choice of recording a single correlation peak at a particular plane. Signal-to-noise ratio and peak-to-sidelobe ratio have been calculated as metrics of goodness of the proposed WaveMACH filter.     

Fingerprint identification by an optical joint transform correlation system

An optical correlator has been used in fingerprint identification, and in order to test its properties in a practical application a compact joint transform correlation system was constructed. The system’s identification performance was qualitatively and statistically evaluated, based on fingerprint data gathered from 500 fellow workers. The power of parallelism and fast processing of optical processing enabled the system to verify an individual’s identity in less than 0.1 s on each try. The system performed a verification with a 4.1% false rejection rate and no false acceptance in “three-try.”