仿射不变的中心投影

 考虑将区域和轮廓类不变特征提取算法结合:利用中心投影将目标转化为闭曲线,对闭曲线运用轮廓类方法提取不变特征.证明了图像经中心投影所得闭曲线保持仿射变换的结论.介绍了一种利用小波变换进行仿射不变特征提取的算法.对一些汉字、26个大写英文字母等的实验结果表明结论的有效性,与传统方法相比该算法对噪声具有较强的鲁棒性.     

基于中心投影的仿射变换参数估计与实现

仿射变换参数恢复轮廓类算法计算量小但一般无法 处理多部分组成 的目标,而区域类算法计算量大且对噪声敏感。本文利用中心投影将区域和轮廓类算法结合 ,首先由中心 投影将目标转化为闭曲线,保证了仿射变换关系具有不变性,再对所得闭曲线进行参数化, 最后利用参数 化后的闭曲线进行仿射变换参数的恢复。实验表明,所提方法可用于多部分组成的目标的参 数恢复,且有较好的参数恢复效果。     

改进的ASIFT算法在图像配准中的应用

SIFT算法具有良好的图像匹配性能,是图像匹配的经典算法之一。针对传统SIFT算法在仿射变换情况下匹配效果不佳、鲁棒性弱的不足,给出一种改进的图像配准方法。首先对图像进行中值滤波处理,去除噪声。然后对图像做仿射变换,提取图像的SIFT特征点进行匹配。最后通过随机抽样一致性方法消除匹配后初始匹配点中的错误匹配点,提高匹配精度。实验结果表明,与SIFT算法相比,提出的方法具有更好的仿射不变性,在平移、旋转、仿射变换情况下均能得到较高的匹配精度,并且有较好的鲁棒性。     

Image registration and object recognition using affine invariantsand convex hulls

This paper is concerned with the problem of feature point registration and scene recognition from images under weak perspective transformations which are well approximated by affine transformations and under possible occlusion and/or appearance of new objects. It presents a set of local absolute affine invariants derived from the convex hull of scattered feature points (e.g., fiducial or marking points, corner points, inflection points, etc.) extracted from the image. The affine invariants are constructed from the areas of the triangles formed by connecting three vertices among a set of four consecutive vertices (quadruplets) of the convex hull, and hence do make direct use of the area invariance property associated with the affine transformation. Because they are locally constructed, they are very well suited to handle the occlusion and/or appearance of new objects. These invariants are used to establish the correspondences between the convex hull vertices of a test image with a reference image in order to undo the affine transformation between them. A point matching approach for recognition follows this. The time complexity for registering L feature points on the test image with N feature points of the reference image is of order O(NxL). The method has been tested on real indoor and outdoor images and performs well.     

图像匹配中具有仿射不变性特征的定量分析

不同视点图像中相应特征点邻域窗口之间存在几何上的透视畸变,这可以用平面单应映射来表示,而目前大多特征匹配算法将该映射用仿射变换模型来近似,即用具有仿射不变性的特征进行图像的匹配。仿射变换的线性特点不仅能降低匹配算法的复杂度,还能保证迭代过程收敛的稳定性,然而并没有人对这一近似的可行性及合理性给出定量的讨论和分析。本文首先回顾各种几何层次上的特征点匹配策略,重点针对具有仿射不变性特征点的定位误差给出定量分析,通过椭圆曲线规范化方法推导出这一近似所造成定位误差的解析表达式,指出用仿射变换模型近似单应映射的合理性;然后用真实图像的实验结果验证了本文分析方法的正确性;最后给出相应的分析结果和结论。     

一种基于独立元分析的仿射不变描述和仿射变换 参数估计的新方法

本文提出了一种基于独立元分析(Independent Component Analysis,ICA)的仿射目标识别和仿射参数估计的新方法.对于待识别的不同扭曲程度的目标,可采用ICA方法提取轮廓的仿射不变描述,从而解决扭曲目标识别问题.该方法同时能够估计两个视频帧间的仿射变换参数,在高压缩率编码标准中,如MPEG4或MPEG7,具有很重要的现实意义.实验结果表明,该方法优于传统的目标识别和仿射参数估计方法.     

Similarity methods in signal processing

A signal may contain information that is preserved by certain transformations of the signal. For example, the information phase-modulated signal is not altered by amplitude scaling of the signal. Many processing techniques have been developed to exploit such similarities. In the past, these algorithms have been developed in isolation without regard to common principles of invariance that tie them together. Similarity methods are presented as a unified method of designing processing algorithms invariant to specified transformations. These methods are based upon groups of continuous transformations known as local Lie groups and lead to a quasilinear partial differential equation. Solution of this partial differential equation specifies the form the signal processing operations must take. This form can then be applied using engineering judgment for algorithmic implementation. The paper presents an extended tutorial on Lie groups and similarity methods and quasilinear differential equations drawn from the mathematical literature. This is followed by several examples of signal processing interest that demonstrate that the similarity techniques may be applicable in certain kinds of signal processing problems     

一种新的仿射不变特征及其在平面曲线匹配中的应用

构造了一种新的仿射变换不变特征,提出了一种适合于仿射变换的曲线匹配算法。首先将整个轮廓区域进行仿射区域划分,得到一系列子区域质心点;然后将质心和子区域的质心点按规则构成三角形,计算三角形的面积比序列,并将其规范化,构造新的仿射不变特征。在该不变特征的基础上设计新的识别向量和差异度度量函数,通过计算待识别目标与模型曲线之...     

Combined affine geometric transformations and spatially dependent radiometric deformations: a decoupled linear estimation framework

This paper considers the problem of registering two observations of the same object, where the observations differ due to a combined effect of an affine geometric transformation and nonuniform illumination changes. The problem of deriving new representations of the observations that are both invariant to geometric transformations and linear in the illumination model is analyzed. In this framework, we present a novel method for linear estimation of illumination changes in an affine invariant manner, thus, decoupling the original problem into two simpler ones. The computational complexity of the method is low as it requires no more than solving a linear set of equations. The prior step of illumination estimation is shown to improve the accuracy of state-of-the-art registration techniques by a factor of two.     

基于点不变特征的亚像素级图像配准技术

图像匹配是图像处理和计算机视觉领域的一个基础问题,它源自多个方面的实际问题.点特征匹配的首要任务就是提取稳定的特征,并进行描述.该特征能对旋转、尺度缩放、仿射变换、视角变化、光照变化等图像变化因素保持一定的不变性,而对物体运动、遮挡、噪声等因素也保持较好的可匹配性.提出一种基于点的尺度、旋转不变特征变换（SIFT）算法的亚像素级图像配准算法,首先在尺度空间进行特征检测,并确定关键点的位置和关键点所处的尺度,然后使用关键点邻域梯度的主方向作为该点的方向特征,以实现算子对尺度和方向的无关性.该方法具有旋转、尺度缩放、亮度变化不变性,对视角变化、仿射变换、噪声也保持一定程度的稳定.独特性好,信息量丰富,适用于在海量特征数据库中进行快速、准确的匹配.高速性,该匹配算法可以达到实时的要求.     

The regularized iteratively reweighted MAD method for change detection in multi- and hyperspectral data.

This paper describes new extensions to the previously published multivariate alteration detection (MAD) method for change detection in bi-temporal, multi- and hypervariate data such as remote sensing imagery. Much like boosting methods often applied in data mining work, the iteratively reweighted (IR) MAD method in a series of iterations places increasing focus on "difficult" observations, here observations whose change status over time is uncertain. The MAD method is based on the established technique of canonical correlation analysis: for the multivariate data acquired at two points in time and covering the same geographical region, we calculate the canonical variates and subtract them from each other. These orthogonal differences contain maximum information on joint change in all variables (spectral bands). The change detected in this fashion is invariant to separate linear (affine) transformations in the originally measured variables at the two points in time, such as 1) changes in gain and offset in the measuring device used to acquire the data, 2) data normalization or calibration schemes that are linear (affine) in the gray values of the original variables, or 3) orthogonal or other affine transformations, such as principal component (PC) or maximum autocorrelation factor (MAF) transformations. The IR-MAD method first calculates ordinary canonical and original MAD variates. In the following iterations we apply different weights to the observations, large weights being assigned to observations that show little change, i.e., for which the sum of squared, standardized MAD variates is small, and small weights being assigned to observations for which the sum is large. Like the original MAD method, the iterative extension is invariant to linear (affine) transformations of the original variables. To stabilize solutions to the (IR-)MAD problem, some form of regularization may be needed. This is especially useful for work on hyperspectral data. This paper describes ordinary two-set canonical correlation analysis, the MAD transformation, the iterative extension, and three regularization schemes. A simple case with real Landsat Thematic Mapper (TM) data at one point in time and (partly) constructed data at the other point in time that demonstrates the superiority of the iterative scheme over the original MAD method is shown. Also, examples with SPOT High Resolution Visible data from an agricultural region in Kenya, and hyperspectral airborne HyMap data from a small rural area in southeastern Germany are given. The latter case demonstrates the need for regularization.     

A multimodal registration algorithm of eye fundus images usingvessels detection and Hough transform

Image registration is a real challenge because physicians handle many images. Temporal registration is necessary in order to follow the various steps of a disease, whereas multimodal registration allows us to improve the identification of some lesions or to compare pieces of information gathered from different sources. This paper presents an algorithm for temporal and/or multimodal registration of retinal images based on point correspondence. As an example, the algorithm has been applied to the registration of fluorescein images (obtained after a fluorescein dye injection) with green images (green filter of a color image). The vascular tree is first detected in each type of images and bifurcation points are labeled with surrounding vessel orientations. An angle-based invariant is then computed in order to give a probability for two points to match. Then a Bayesian Hough transform is used to sort the transformations with their respective likelihoods. A precise affine estimate is finally computed for most likely transformations. The best transformation is chosen for registration.     

仿射不变子空间特征及其在图像匹配中的应用

提出了一种新的基于子空间的仿射不变图像匹配算法。首先,分析了仿射不变特征空间的非线性几何结构,给出了图像特征的仿射不变子空间表示模型。其次,根据目标特征矩阵由它的正交投影矩阵唯一确定这一特性实现特征的全局一致性匹配。最后,通过计算同一目标在仿射变换下的子空间几何距离度量达到匹配识别的目的。模拟变换匹配实验和真实图像序列实验结果表明:所提出的匹配方法对仿射变化下的图像目标匹配具有很好的适应性和较高的鲁棒性。     

仿射不变三角形在红外目标识别中的应用

红外目标识别主要存在以下两个问题：一是红外成像比较模糊,难以提取不变特征;二是红外图像信息含量小,不易发掘有效的描述方法。针对上述问题,提出了一种基于曲率尺度空间(Curvature Scale Space,CSS)角点的仿射不变三角形的检测方法,并利用优选参数的多尺度自卷积(Multi-scale Autoconvolution,MSA)对提取的三角形区域进行描述,最后进行特征匹配实现红外目标识别。实验表明：与其他方法比较,在各种图像变换中,本文方法对红外目标的识别优势显著。     

Comparison of GLR and invariant detectors under structured cluttercovariance

This paper addresses a target detection problem in radar imaging for which the covariance matrix of unknown Gaussian clutter has block diagonal structure. This block diagonal structure is the consequence of a target lying along a boundary between two statistically independent clutter regions. Here, we design adaptive detection algorithms using both the generalized likelihood ratio (GLR) and the invariance principles. There has been considerable interest in applying invariant hypothesis testing as an alternative to the GLR test. This interest has been motivated by several attractive properties of invariant tests including: exact robustness to variation of nuisance parameters and possible finite-sample min-max optimality. However, in our deep-hide target detection problem, there are regimes for which neither the GLR nor the invariant tests uniformly outperforms the other. We discuss the relative advantages of GLR and invariance procedures in the context of this radar imaging and target detection application.