一类新的正交矩-Franklin矩及其图像表达

该文定义了一类以Franklin函数为核的正交矩,称之为Franklin矩.Franklin函数是一类完备正交一次样条函数系.传统的Legendre矩、Zernike矩等多项式矩,由于涉及高次多项式的计算,往往会导致计算不稳定,特征空间维数扩展受到制约.Franklin函数是正交的,相应的矩函数可以使得图像分解后的信息具有独立性,没有信息的冗余.而且,Franklin函数仅由一次分段多项式组成,在计算过程中,避免了高次多项式的计算,兼具复杂度低、数值稳定的优点.通过对图像的重构实验表明,Franklin矩比传统正交多项式矩具有更好的特征表达能力.     

二维旋转不变U变换及其应用

U-系统是一类L2[0,1]上的正交分段多项式函数系,为了将其推广到二维情形,传统的L2[0,1]2上张量积形式的U变换并不具有旋转不变性.本文提出了一类二维旋转不变U变换（Rotation-invariant U transform,RIUT）. RIUT将U-系统函数与调和函数相结合,使得图像的旋转转化为相位的平移而模保持不变.与经典的正交旋转不变矩（如Zernike矩）相比,RIUT具有诸多特别的性质,从而在图像特征提取中具有良好的潜力.本文将RIUT应用到二值图像检索中的实验结果表明,RIUT具有更高的检索精度.     

基于两变量Krawtchouk矩的图像重建

传统的离散正交Krawtchouk矩的基函数由两个单变量的Krawtchouk多项式乘积构成,它割裂平面两个方向之间的联系。提出了一种新的、以两变量Krawtchouk正交多项式为基函数的图像矩,并推导了正则化后两变量多项式的简单的计算方法。重建实验结果表明,相对于同系数的单变量的离散正交矩,两变量离散正交矩的重建误差更小。     

参数可调的通用半正交图像矩模型

目的 为了提高以正交多项式为核函数构造的高阶矩数值的稳定性,增强低阶矩抗噪和滤波的能力,将仅具有全局描述能力的常规正交矩推广到可以局部化提取图像特征的矩模型,从频率特性分析的角度定义一种参数可调的通用半正交矩模型。方法 首先,对传统正交矩的核函数进行合理的修正,以修正后的核函数（也称基函数）替代传统正交矩中的原核函数,使其成为修改后的特例之一。经过修正后的基函数可以有效消除图像矩数值不稳定现象。其次,采用时域的分析方法能够对图像的低阶矩作定量的分析,但无法对图像的高频部分（对应的高阶矩）作更合理的表述。因此提出一种时—频对应的方法来分析和增强不同阶矩的稳定性,通过对修正后核函数的频带宽度微调可以建立性能更优的不同阶矩。最后,利用构建的半正交—三角函数矩研究和分析了通用半正交矩模型的特点及性质。结果 将三角函数为核函数的图像矩与现有的Zernike、伪Zernike、正交傅里叶—梅林矩及贝塞尔—傅里叶矩相比,由于核函数组成简单,且其值域恒定在[-1,1]区间,因此在图像识别领域具有更快的计算速度和更高的稳定性。结论 理论分析和一系列相关图像的仿真实验表明,与传统的正交矩相比,在数值稳定性、图像重构、图像感兴趣区域（ROI）特征检测、噪声鲁棒性测试及不变性识别方面,通用的半正交矩性能及效果更优。     

形状和颜色变换下图像的Gaussian-Hermite矩不变量

矩不变量是计算机视觉和模式识别领域常用的图像不变特征.现有的形状和颜色变换下彩色图像的矩不变量均基于几何矩构造,因此抗噪性较差.针对该问题,提出了利用基本微分算子和颜色几何基元生成旋转-仿射变换下彩色图像Gaussiaa-Hermite正交矩不变量的方法,所构造的不变量均为Gaussian-Hermite矩的齐次多项式...     

抗几何攻击图像水印研究

研究图像数字水印安全问题.抵抗几何攻击一直是水印算法研究的难点和重点,在图像版权保护过程的研究,针对旋转、缩放和平移等几何攻击能够破坏水印检测的同步性,容易使常规水印算法检测不稳定的问题,利用图像伪Zernike矩的幅度具有旋转不变的性质,提出了一种可有效抵抗几何攻击的数字图像水印新方法.利用图像归一化技术将原始载体映射到几何不变空间内,以消除缩放、平移之影响,然后结合伪Zernike矩的旋转不变特性,计算出归一化图像的伪Zernike矩.最后选取部分低阶伪Zernike矩,并采纳量化调制策略将水印信息嵌入到伪Zernike矩幅值中.仿真结果表明,数字图像水印算法不仅具有良好的透明性,而且具有较强的抗攻击能力,为设计提供了可靠依据.     

新形式勒让德矩的研究

针对传统非正交矩很难进行图像重建的缺点,以及离散矩用于重建需要重复采样的缺陷,以降低图像重建误差为目标,提出了一种以在离散坐标空间内拟合克罗内克狄拉克函数为核心思想的新形式矩的定义--基于勒让德多项式的矩,并对其性质进行了阐述。这种矩在函数空间非正交却拥有优秀的重建效果,且其在矩计算误差、旋转不变性等多个维度较目前主流矩都具有更优秀性能,特别是在目前主流图像矩表现不尽如人意的大尺寸图像领域。此外,突破性地发掘图像矩的抗噪音性能并加入性能对比。通过与目前主流的三种矩：Zernike矩、Polar-Fourier矩以及Polar Harmonic Transform（PHT）矩的对比实验,证明利用这种基于新思想的矩提取图像特征可以具有更小的信息冗余度及多个维度的鲁棒性,其在旋转不变性、减小图像重建误差以及提高抗噪稳定性方面的性能表现至少可以提高22%。     

基于两变量Hahn矩的图像分析

提出了一种新的、以两变量离散正交Hahn多项式为核函数的图像矩,推导了正则化后,两变量离散正交Hahn多项式的简单的计算方法。对二值图像、灰度图像以及噪声图像的重建实验表明：相对于同系数的单变量的Hahn矩,两变量Hahn矩的重建误差更小。因此,它们能够更好地提取图像的特征。     

基于三角函数的正交矩函数的误差预测与分析

提取图像中旋转不变特征是图像处理和模式识别中重要的应用。在极坐标下的正交矩函数则是提取这种特征信息的主要方法。正交矩函数在图像分解和重建过程中的误差是衡量其特征提取精确度的标准。为了提高正交矩函数在图像重建中的性能,提出了一种新的基于三角函数的正交矩函数和一种基于函数误差分析的新的衡量方法,并分别应用新的衡量方法和传统的在大量图像中进行重建误差统计的方法对新的正交矩函数以及另外两种在特征提取方面表现最好的正交矩函数进行了比较。实验结果验证了新的衡量方法的有效性并得到了新的正交矩函数的重建效果更好的结论。     

Zernike亚像素图像矩边缘检测算子方向角模型研究

基于Zernike图像矩的理想边缘模型,深入研究了方向角模型与亚像素判据间的关系。利用Zernike矩定义及其旋转不变特性,提出一种新的基于4阶方向角的Zernike矩亚像素边缘检测算子。为了提高边缘算子定位速度,首先基于9×9尺寸模板对Zernike图像矩0～4阶正交复数多项式进行了计算,推导出基于4阶方向角的边缘检测算子参数模型。最后将边缘算子应用在理想图像与实际图像上,检测结果表明:相比于传统的Zernike矩算子,基于4阶方向角的边缘检测算子具有更高的检测精度。     

An efficient and robust multi-frame image super-resolution reconstruction using orthogonal Fourier-Mellin moments

Multi-frame image super-resolution (SR) has recently become an active area of research. The orthogonal rotation invariant moments (ORIMs) have several useful characteristics which make them very suitable for multi-frame image super-resolution application. Among the various ORIMs, Zernike moments (ZMs) and pseudo-Zernike moments (PZMs)-based SR approaches, i.e., NLM-ZMs and NLM-PZMs, have already shown improved SR performances for multi-frame image super-resolution. However, it is a well-known fact that among many ORIMs, orthogonal Fourier-Mellin moments (OFMMs) demonstrate better noise robustness and image representation capabilities for small images as compared to ZMs and PZMs. Therefore, in this paper, we propose a multi-frame image super-resolution approach using OFMMs. The proposed approach is based on the NLM framework because of its inherent capability of estimating motion implicitly. We have referred to this proposed approach as NLM-OFMMs-I. Also, a novel idea of using OFMMs-based interpolation in place of traditional Lanczos interpolation for obtaining an initial estimate of HR sequence has been presented in this paper. This variant of the proposed approach is referred to as NLM-OFMMs-II. Detailed experimental analysis demonstrates the effectiveness of the proposed OFMMs-based SR approaches to generate high-quality HR images in the presence of factors like image noise, global motion, local motion, and rotation in between the image frames.     

Rotation-invariant texture image retrieval using particle swarm optimization and support vector regression

This paper presents a novel rotation-invariant texture image retrieval using particle swarm optimization (PSO) and support vector regression (SVR), which is called the RTIRPS method. It respectively employs log-polar mapping (LPM) combined with fast Fourier transformation (FFT), Gabor filter, and Zernike moment to extract three kinds of rotation-invariant features from gray-level images. Subsequently, the PSO algorithm is utilized to optimize the RTIRPS method. Experimental results demonstrate that the RTIRPS method can achieve satisfying results and outperform the existing well-known rotation-invariant image retrieval methods under considerations here. Also, in order to reduce calculation complexity for image feature matching, the RTIRPS method employs the SVR to construct an efficient scheme for the image retrieval.     

Invariant image recognition by Zernike moments

The problem of rotation-, scale-, and translation-invariant recognition of images is discussed. A set of rotation-invariant features are introduced. They are the magnitudes of a set of orthogonal complex moments of the image known as Zernike moments. Scale and translation invariance are obtained by first normalizing the image with respect to these parameters using its regular geometrical moments. A systematic reconstruction-based method for deciding the highest-order Zernike moments required in a classification problem is developed. The quality of the reconstructed image is examined through its comparison to the original one. The orthogonality property of the Zernike moments, which simplifies the process of image reconstruction, make the suggest feature selection approach practical. Features of each order can also be weighted according to their contribution to the reconstruction process. The superiority of Zernike moment features over regular moments and moment invariants was experimentally verified     

VirtMan: design and implementation of a fast booting system for homogeneous virtual machines in iVCE

Internet-based virtual computing environment (iVCE) has been proposed to combine data centers and other kinds of computing resources on the Internet to provide efficient and economical services. Virtual machines (VMs) have been widely used in iVCE to isolate different users/jobs and ensure trustworthiness, but traditionally VMs require a long period of time for booting, which cannot meet the requirement of iVCE’s large-scale and highly dynamic applications. To address this problem, in this paper we design and implement VirtMan, a fast booting system for a large number of virtual machines in iVCE. VirtMan uses the Linux Small Computer System Interface (SCSI) target to remotely mount to the source image in a scalable hierarchy, and leverages the homogeneity of a set of VMs to transfer only necessary image data at runtime. We have implemented VirtMan both as a standalone system and for OpenStack. In our 100-server testbed, VirtMan boots up 1000 VMs (with a 15 GB image of Windows Server 2008) on 100 physical servers in less than 120 s, which is three orders of magnitude lower than current public clouds.     

A moment-based nonlocal-means algorithm for image denoising

Image denoising is a crucial step to increase image quality and to improve the performance of all the tasks needed for quantitative imaging analysis. The nonlocal (NL) means filter is a very successful technique for denoising textured images. However, this algorithm is only defined up to translation without considering the orientation and scale for each image patch. In this paper, we introduce the Zernike moments into NL-means filter, which are the magnitudes of a set of orthogonal complex moments of the image. The Zernike moments in small local windows of each pixel in the image are computed to obtain the local structure information for each patch, and then the similarities according to this information are computed instead of pixel intensity. For the rotation invariant of the Zernike moments, we can get much more pixels or patches with higher similarity measure and make the similarity of patches translation-invariant and rotation-invariant. The proposed algorithm is demonstrated on real images corrupted by white Gaussian noise (WGN). The comparative experimental results show that the improved NL-means filter achieves better denoising performance.