基于四元数傅里叶梅林变换的旋转不变彩色纹理分类

提出一种基于四元数傅里叶梅林变换(Quaternion Fourier-Mellin Transform,QFMT)的旋转不变彩色纹理分类方法。该方法首先对彩色图像各分量图像进行对数极坐标变换,然后将经过变换后的3幅分量图像表示成四元数,并对其进行四元数傅里叶变换(Quaternion Fourier Transform,QFT),最后对幅度谱分别统计其环形特征量和楔形特征量作为纹理分类的特征向量,利用最近邻分类器进行分类。实验结果表明,本文提出的方法分类准确率更高,且具有良好的旋转不变纹理分析性能。     

基于角度径向变换的旋转不变纹理分类

纹理分类广泛的应用于医学图像分析等领域,纹理图像的采集因拍摄角度的变化产生一定的旋转,本文提出一种基于角度径向变换的旋转不变纹理分类方法。首先采用角度径向变换方法对图像进行特征提取,分别得到图像的角向特征向量和径向特征向量;然后将提取出的2组特征向量结合起来作为图像的整体特征向量,利用K近邻特征空间距离的分类方法进行纹理分类。选取Brodatz纹理库中的图像进行纹理分类测试,实验结果表明,该算法具有较好的旋转不变纹理分类效果。     

基于旋转不变纹理特征的多尺度多方向图像渐进检索

纹理检索是基于内容图像检索的重要内容,旋转不变纹理图像检索是实现纹理检索的关键途径之一.针对旋转不变纹理图像检索中需要解决的3个关键问题:如何消除旋转影响、如何选择多尺度分析方法以及如何构造和度量纹理特征矢量,本文分别分析了Radon变换和Log-polar变换在消除旋转位移时对频谱的影响,以及NSCT变换和小波变换在不同检索参数下的平均检索性能,在此基础上构造出多尺度多方向纹理变换谱和旋转不变特征矢量,提出一种多尺度多方向旋转不变纹理图像渐进检索方法.这种方法采用了可顾及人类视觉对纹理能量敏感性的相似性度量标准,分别采用旋转位移处理后的NSCT变换域低频子带和高通子带实现纹理图像的粗检索和精细检索.Brodatz标准纹理图像库的检索实验表明,本文提出的利用多尺度多方向纹理变换谱构造旋转不变特征矢量的方法既可获取纹理主方向,同时又能有效地表征纹理细节信息,两级渐进式检索策略与多尺度分析方法相结合,既能提高旋转不变纹理图像检索的查准率,又能保证较高的检索效率.     

基于Radon变换和SWT的旋转不变纹理分类

提出了一种应用Radon变换和离散平稳小波变换(SWT)的旋转不变纹理分类算法。该方法首先对纹理图像进行Radon变换将旋转转化为平移,再用具有平移不变性的离散平稳小波对变换后的图像滤波并计算各子带的能量值组成旋转不变特征向量,最后利用支持向量机实现纹理图像的分类。将本方法与其它旋转不变纹理分类法进行比较,实验结果表明,提出的方法能有效地提高正确分类率。     

基于Radon变换和EMD的旋转不变纹理分类"

提出一种结合Radon变换和经验模态分解(EMD)的旋转不变纹理分类方法。对纹理图像进行Radon变换,使用EMD分解成一系列固有模态函数(IMF),用修改过的能量公式计算各IMF的能量并组成特征向量,利用k-近邻对纹理图像进行分类。通过对Brodatz纹理库中的图像进行分类实验,表明所提方法能够有效地进行旋转不变的纹理分类,较之单一的Radon变换分类有了显著的提高。     

基于Log-Polar和DT-CWT的旋转不变纹理分类算法

提出了一种基于对数-极坐标变换和双树复数小波变换的旋转不变纹理分类算法。该方法首先对纹理图像进行对数-极坐标变换将旋转转化为平移,再用具有平移不变性的双树复数小波对变换后的图像滤波并计算各子带的能量值组成旋转不变特征向量,最后利用支持向量机算法实现纹理图像的分类。将本方法与其它旋转不变纹理分类算法进行比较,实验结果表明,提出的算法能有效地提高正确分类率。     

基于纹理的旋转不变图像检索算法的研究

提出一种基于Gabor变换的旋转不变多尺度广义粗糙度特征向量并结合自适应加权距离进行纹理图像检索的方法。利用图像Gabor分解的幅度谱,依据多尺度空间局部能量分布、Hurst分形指数、方向差别来计算纹理特征向量,最后采用自适应加权的街区距离作为相似性准则。仿真结果表明,该算法对旋转纹理图像取得了很好的检索结果。     

基于尺度相关性的旋转不变纹理图像渐进式检索算法

在常用的基于小波变换域旋转不变纹理图像检索算法中,由于存在方向信息提取有限且多尺度间系数相关性被忽略的局限性,检索效率受到影响。提出一种基于尺度相关性的渐进式旋转不变纹理图像检索算法。该算法首先采用Log-polar变换与非下采样Contourlet变换组合的方式获取具备旋转不变性的多尺度多方向变换系数,然后利用广义高斯模型拟合低通波段的全局结构信息作为粗判依据,方向子带间的尺度相关信息则采用非高斯双变量模型拟合,并作为精细渐进式检索的特征变量。基于Brodatz标准纹理库的实验结果表明,与小波变换及基于层内关系模型方法相比,该方法能以更低的特征维数获得更高的检索效率及检索准确率,是一种进行旋转纹理检索的有效手段。     

旋转、缩放、位移不变的小波域图像水印算法

介绍了一种基本上不受旋转、缩放、位移影响的小波域图像水印算法.该算法根据人类视觉系统的特性,将水印自适应地嵌入图像小波域的低频子带.为计算和校正旋转及缩放造成水印位置的移动,采用一种新的滤波方法.该滤波方法需要使用原始图像对数极坐标域中的一小方块信息来计算水印位置的平移.实验结果表明,该水印对旋转、缩放、平移和常规的图像处理具有较强的鲁棒性.     

完备局部凸凹计数模式及其旋转不变纹理分类

局部凸凹模式是新近提出的一种有效纹理描述子,但对旋转纹理图像描述不足。针对局部凸凹模式算法的此缺陷,提出了局部凸凹计数模式进行旋转不变纹理描述。局部凸凹计数模式统计纹理的局部凸或凹模式数目,而抛弃了局部凸凹模式的结构信息。在纹理图像发生平面内旋转情况下,局部凸凹计数模式所提取的纹理特征具有不变性。另外,为了增强局部凸凹计数模式的性能,提出了完备局部凸凹计数模式算法。完备局部凸凹计数模式不仅统计了局部凸或凹模式数目,还对局部纹理的凸或凹强度进行了描述,中心像素所包含的鉴别信息也统计在完备局部凸凹模式算法中。二个纹理数据库和一个掌纹数据库上的试验充分表明完备局部凸凹计数模式是一种有效的旋转不变纹理特征提取算法。     

Image segmentation from scale and rotation invariant texture features from the double dyadic dual-tree complex wavelet transform

A goal of image segmentation is to divide an image into regions that have some semantic meaning. Because regions of semantic meaning often include variations in colour and intensity, various segmentation algorithms that use multi-pixel textures have been developed. A challenge for these algorithms is to incorporate invariance to rotation and changes in scale. In this paper, we propose a new scale and rotation invariant, texture-based segmentation algorithm, that performs feature extraction using the Dual-Tree Complex Wavelet Transform (DT-CWT). The DT-CWT is used to analyse a signal at, and between, dyadic scales. The performance of image segmentation using this new method is compared with existing techniques over different imagery databases with operator produced ground truth data. Compared with previous algorithms, our segmentation results show that the new texture feature is capable of performing well over general images and particularly well over images containing objects with scaled and rotated textures.     

The multiple-parameter fractional Fourier transform

The fractional Fourier transform （FRFT） has multiplicity, which is intrinsic in fractional operator. A new source for the multiplicity of the weight-type fractional Fourier transform （WFRFT） is proposed, which can generalize the weight coefficients of WFRFT to contain two vector parameters m,n ∈ Z^M . Therefore a generalized fractional Fourier transform can be defined, which is denoted by the multiple-parameter fractional Fourier transform （MPFRFT）. It enlarges the multiplicity of the FRFT, which not only includes the conventional FRFT and general multi-fractional Fourier transform as special cases, but also introduces new fractional Fourier transforms. It provides a unified framework for the FRFT, and the method is also available for fractionalizing other linear operators. In addition, numerical simulations of the MPFRFT on the Hermite-Gaussian and rectangular functions have been performed as a simple application of MPFRFT to signal processing.     

Recursive sliding discrete Fourier transform with oversampled data

The Discrete Fourier Transform (DFT) has played a fundamental role for signal analysis. A common application is, for example, an FFT to compute a spectral decomposition, in a block by block fashion. However, using a recursive, discrete, Fourier transform technique enables sample-by-sample updating, which, in turn, allows for the computation of a fine time–frequency resolution. An existing spectral output is updated in a sample-by-sample fashion using a combination of the Fourier time shift property and the difference between the most recent input sample and outgoing sample when using a window of finite length. To maintain sampling-to-processing synchronisation, a sampling constraint is enforced on the front–end hardware, as the processing latency per input sample will determine the maximum sampling rate. This work takes the recursive approach one step further, and enables the processing of multiple samples acquired through oversampling, to update the spectral output. This work shows that it is possible to compute a fine-grained spectral decomposition while increasing usable signal bandwidths through higher sampling rates. Results show that processing overhead increases sub-linearly, with signal bandwidth improvement factors of up to 6.7× when processing 8 samples per iteration.     

基于NSCT的旋转不变纹理图像检索算法

针对图像检索中常见的旋转问题,提出了一种基于非抽样Contourlet变换(NSCT)的旋转不变检索算法。以NSCT域各子带系数的均值和标准方差构成基本特征向量。在相同尺 度上,利用各子带的均值和标准方差之和对特征分量由小到大排序,同时根据旋转不变性调整排序后特征向量对应的方向序列,构造方向序列权值、特征分量权值。用加权欧氏距 离进行相似性度量以提高检索性能。采用Brodatz库生成实验库,实验结果表明本文方法取得了较好的检索效果。     

Research progress on discretization of fractional Fourier transform

As the fractional Fourier transform has attracted a considerable amount of attention in the area of optics and signal processing, the discretization of the fractional Fourier transform becomes vital for the application of the fractional Fourier transform. Since the discretization of the fractional Fourier transform cannot be obtained by directly sampling in time domain and the fractional Fourier domain, the discretization of the fractional Fourier transform has been investigated recently. A summary of discretizations of the fractional Fourier transform developed in the last nearly two decades is presented in this paper. The discretizations include sampling in the fractional Fourier domain, discrete-time fractional Fourier transform, fractional Fourier series, discrete fractional Fourier transform （including 3 main types： linear combination-type; sampling-type; and eigen decomposition-type）, and other discrete fractional signal transform. It is hoped to offer a doorstep for the readers who are interested in the fractional Fourier transform.