基于自适应滤波的快速广义模糊C均值聚类图像分割

快速广义模糊C均值聚类(FGFCM)在对高噪声图像进行聚类分割时,噪声容易导致聚类中心发生偏移,影响图像分割结果.为此,文中提出基于自适应滤波的快速广义模糊C均值聚类算法,用于图像分割.首先根据非局部像素的噪声概率自适应确定参数平衡因子,更准确地反映图像包含的空间结构信息.然后利用该平衡因子有效结合FGFCM中的线性加权和滤波图像与原始图像的中值滤波图像,由于得到的自适应滤波图像根据图像中像素为噪声的概率自适应确定滤波程度,因此可以提高算法对噪声的动态抑制能力.实验表明,相比模糊C均值聚类和FGFCM,文中算法在对噪声含量较高的图像进行聚类分割时,可以得到更准确的结果.     

基于空间特征的谱聚类含噪图像分割

为克服传统谱聚类算法应用到含噪图像分割时易受到图像中噪声影响的问题,提出一种基于空间特征的谱聚类含噪图像分割算法。该方法利用图像各个像素的灰度信息、局部空间邻接信息及非局部空间信息设计像素的三维特征,通过引入空间紧致性函数建立像素特征点与其K个最近邻之间的相似性,进而利用谱聚类算法得到图像的最终分割结果。实验中采用含噪的人工图像、自然图像及合成孔径雷达图像与空间模糊聚类、规范切谱聚类和Nystrm方法3种算法进行对比实验,实验结果验证文中方法能克服图像中噪声影响并取得较满意的分割效果。     

基于目标区域的语义图像检索

提出了一种基于高层语义的图像检索方法,该方法首先将图像分割成区域,提取每个区域的颜色、形状、位置特征,然后使用这些特征对图像对象进行聚类,得到每幅图像的语义特征向量;采用模糊C均值算法对图像进行聚类,在图像检索时,查询图像和聚类中心比较,然后在距离最小的类中进行检索。实验表明,提出的方法可以明显提高检索效率,缩小低层特征和高层语义之间的“语义鸿沟”。     

Stability and dimensionality of Karhunen-Loêve multispectral image expansions

The Karhunen-Loêve (K.L.) expansion is a useful tool for the representation, pre-processing and orthogonal coding of multispectral imagery: Each spatial pixel is analysed independently as the K.L. transform is taken in the spectral dimension, i.e. along the various N spectral channels. The eigenvectors are those of the covariance matrix. The (principal) eigenimages are thus “false color” images, which can be viewed without decoding as the spatial topology is unchanged, and the higher order principal images present a strong contrast enhancement.⁽¹⁾ These principal images are also uncorrelated, a very desirable feature for many applications including clustering.⁽²⁾

The source dependency of the eigenvectors, however, introduces “instability” in the form of pronounced statistical noise on some principal images. This paper gives the results of a numerical study carried out on a 7 channel Daedalus Multispectral Scene. The uncertainties of the eigenvalues and eigenvectors are evaluated from two “drawings” of the pixels of the raw data.

Both the numerical results of the study and the direct viewing of the principal images show that three out of the seven have so much noise that they do not yield any useful information. Only the first two principal images have excellent stability, and they contain most of the total contrast variance of the scene. Two other principal images of lower order are also stable, but, contribute very little to the total contrast variance. These images carry texture information rather than homogeneous zone clustering information.     

DBCAMM: A novel density based clustering algorithm via using the Mahalanobis metric

In this paper we propose a new density based clustering algorithm via using the Mahalanobis metric. This is motivated by the current state-of-the-art density clustering algorithm DBSCAN and some fuzzy clustering algorithms. There are two novelties for the proposed algorithm: One is to adopt the Mahalanobis metric as distance measurement instead of the Euclidean distance in DBSCAN and the other is its effective merging approach for leaders and followers defined in this paper. This Mahalanobis metric is closely associated with dataset distribution. In order to overcome the unique density issue in DBSCAN, we propose an approach to merge the sub-clusters by using the local sub-cluster density information. Eventually we show how to automatically and efficiently extract not only ‘traditional’ clustering information, such as representative points, but also the intrinsic clustering structure. Extensive experiments on some synthetic datasets show the validity of the proposed algorithm. Further the segmentation results on some typical images by using the proposed algorithm and DBSCAN are presented in this paper and they are shown that the proposed algorithm can produce much better visual results in image segmentation.     

结合纹理去除的遥感图像分割

针对包含复杂纹理信息的遥感图像难以进行精准图像分割的问题,提出了一种结合纹理去除的遥感图像分割方法。首先,改进了相对全变差纹理去除方法,通过引入新的范数约束使相对全变差纹理去除方法可以在去除纹理信息的同时凸显图像中的主要结构,达到辅助分割的效果;然后,使用均值漂移算法对经过纹理去除的遥感图像进行无监督聚类,达到分割的目的;最后,提出的遥感图像分割算法在不同遥感图像上进行了测试。实验结果表明,在高分辨遥感图像的分割上,所提算法可以分割出遥感图像中的主要目标,和直接分割或者结合其他纹理去除方法相比取得了更好的分割结果。所提出的分割算法可以降低纹理信息对图像分割的影响,提高遥感图像分割的精度。     

快速鲁棒核空间模糊聚类分割

目的 传统模糊C-均值聚类应用于图像分割仅考虑像素本身的聚类问题,无法克服噪声干扰对图像分割结果的影响,不利于受到噪声干扰的工业图像、医学影像和高分遥感影像等进行目标提取、识别和解译。嵌入像素空间邻域信息或局部信息的鲁棒模糊C-均值聚类分割算法是近年来图像分割理论研究中的热点课题。为此,针对现有的鲁棒核空间模糊聚类算法非常耗时且抑制噪声能力弱、不适合强噪声干扰下大幅面图像快速分割等问题,提出一种快速鲁棒核空间模糊聚类分割算法。方法 利用待分割图像中像素邻域的灰度信息和空间位置等信息构建线性加权滤波图像,对其进行鲁棒核空间模糊聚类。为了进一步提高算法实时性,引入当前聚类像素与其邻域像素均值所对应的2维直方图信息,构造一种基于2维直方图的鲁棒核空间模糊聚类快速分割最优化数学模型,采用拉格朗日乘子法获得图像分割的像素聚类迭代表达式。结果 对大幅面图像添加一定强度的高斯、椒盐以及混合噪声,以及未加噪标准图像的分割测试结果表明,本文算法比基于邻域空间约束的核模糊C-均值聚类等算法的峰值信噪比至少提高1.5 dB,误分率降低约5%,聚类性能评价的划分系数提高约10%,运行速度比核模糊C-均值聚类和基于邻域空间约束的鲁棒核模糊C-均值聚类算法至少提高30%,与1维直方图核空间模糊C-均值聚类算法具有相当的时间开销,所得分割结果具有较好的主观视觉效果。结论 通过理论分析和实验验证,本文算法相比现有空间邻域信息约束的鲁棒核空间模糊聚类等算法具有更强的抗噪鲁棒性、更优的分割性能和实时性,对大幅面遥感、医学等影像快速解译具有积极的促进作用,能更好地满足实时性要求较高场合的图像分割需要。     

一种新的基于语义聚类和图算法的自动图像标注方法

针对图像检索中的语义鸿沟问题,提出了一种新颖的自动图像标注方法。该方法首先采用了一种基于软约束的半监督图像聚类算法(SHMRF-Kmeans)对已标注图像的区域进行语义聚类,这种聚类方法可以同时考虑图像的视觉信息和语义信息。并利用图算法——Manifold排序学习算法充分发掘语义概念与区域聚类中心的关系,得到两者的联合概率关系表。然后利用此概率关系表标注未知标注的图像。该方法与以前的方法相比可以更加充分地结合图像的视觉特征和高层语义。通过在通用图像集上的实验结果表明,本文提出的自动图像标注方法是有效的。     

鲁棒空间约束的模糊聚类图像分割

目的 为进一步提高分割精度,在模糊聚类的基础上引入统计信息,提出一种鲁棒型空间约束的模糊聚类分割算法。方法 基于局部空间信息的先验概率与后验概率,提出一种新型空间约束项,并通过卷积操作提高运行效率;进而引入负对数联合概率作为测度函数,进一步提高算法对于各像素点所属类别的甄别能力;同时将测度函数与空间约束项整合至目标函数中,通过迭代更新各参数达到最小化目标函数的目的。结果 对于合成图像的实验结果表明,本文算法对于噪声类型和噪声强度具有较强的鲁棒性;对于彩色图像的实验结果表明,在适当的特征描述符的辅助下,本文算法也能够获得令人满意的分割结果和较高的分割精度。结论 本文算法克服了现有算法的缺陷,进一步提升了图像的分割精度。其适用于分割带噪声图像,且在适当纹理特征的辅助下分割彩色图像,与同类算法的比较实验结果验证了本文算法的有效性。     

显著特征融合的主颜色聚类分割算法*

针对颜色密度聚类分割模型容易产生误分割的问题,提出基于视觉显著性调节的主颜色聚类分割算法.首先,根据空间颜色信息和Mean-shift算法平滑结果分别计算图像的全局显著特征和区域显著特征,并融合2类显著特征作为特征空间聚类的约束项.然后,采用核密度估计方法计算图像主颜色作为初始类,并将显著特征作为调节因子进行聚类分割.最后,进行区域合并.在标准的分割图像库上进行实验并与多种算法对比,结果表明,文中算法具有更高的区域轮廓准确度,并且有效利用图像显著特征,降低密度聚类形成的区域不一致性,提高像素聚类的精度和分割的鲁棒性.     

Expert vision systems integrating image segmentation and recognition processes

In this paper, we propose a prototype rule-based system which integrates segmentation and recognition processes to analyze and classify objects in an image. This is quite different from the traditional image analysis paradigm which treats segmentation as a prerequisite for recognition and interpretation. There are four basic components in the system, i.e., low-level image processing, feature computation, domain-independent, and domain-dependent subsystems. In the low-level image processing subsystem, various “nonpurposive” operators are employed to divide the image into uniform and homogeneous regions based on the information of intensities. The feature computation subsystem extracts features of each individual region. The domain-independent subsystem employs weak knowledge to filter out “obviously impossible” regions while the domain-dependent subsystem uses domain-specific knowledge to improve the results and finally recognize the objects of interest in the image. Two sets of images are used to demonstrate the capability and flexibility of this system. One set consists of distributor caps (auto parts) of different shapes. The other set is composed of tomographical image pairs acquired by MRI and PET.     

结合分水岭的纹理梯度各向异性图像分割

目的 目前,许多图像分割算法对含有丰富纹理信息的图像的分割效果并不理想,尤其是在不同纹理的边缘信息的保持方面。为了解决这一问题,提出一种基于连续纹理梯度信息的各向异性图像分割算法。方法 在分水岭算法的基础上,引入纹理梯度各向异性算法,能够在避免纹理信息影响分割效果的前提下,最大限度地保证纹理边缘信息的完整。针对纹理特征数据敏感的特性,本文将离散的图像高度信息映射到连续的纹理梯度空间,能够有效减少由细小差异造成的过分割现象。结果 本文方法在BSD500 Dataset和Stanford Background Dataset中选择了大量的纹理信息丰富的图片与最新的分割算法进行了实验与对比。本文方法在分割效果（降低过分割现象）、保持边缘信息和分割准确率等方面均获得明显改进,并在图像分割的平均准确率方面与最新算法进行比较发现,本文算法的平均分割准确率达到90.9%,明显超过了其他最新算法,验证了本文方法的有效性。结论 本文提出的基于分水岭的纹理梯度各向异性算法对纹理图像的分割具有保边和准确的特点,采用连续梯度空间的方法能够有效地减少传统分水岭算法的过分割现象。本文方法主要适用于纹理信息丰富（自然纹理和人工纹理）的图片。     

基于免疫谱聚类的图像分割

提出了一种基于免疫谱聚类的图像分割方法.利用谱聚类的维数缩减特性获得数据在映射空间的分布,在此基础上构造一种新的免疫克隆聚类,用于在映射空间中对样本进行聚类.该方法通过谱映射为后续的免疫克隆聚类提供低维而紧致的输入.而免疫克隆聚类算法具有快速收敛到全局最优并且对初始化不敏感的特性,从而可以获得良好的聚类结果.在将其用于图像分割时,采用了Nystr?m逼近策略来降低算法复杂度.合成纹理图像和SAR图像的分割结果验证了免疫谱聚类算法用于图像分割的有效性.     

基于有偏场的光栅图像模糊聚类分割算法

针对有偏场环境下带有光栅的散焦图像分割问题,提出了一种新的基于有偏场估计的模糊聚类分割算法。通过建立依赖于有偏场的模糊聚类目标函数,导出了基于灰度以及邻域灰度均值的聚类中心、模糊聚类函数以及有偏场估计的迭代算法;并在该算法生成的初始分割基础上,利用膨胀算子对分类结果进行细化。该方法较好地处理了传统模糊聚类对有偏场下光栅图像分割精度下降的问题。实验结果表明,基于有偏场的模糊聚类算法能有效分割光栅图像,其分割精度优于传统模糊聚类和阈值法。     

A local-spectral fuzzy segmentation for MSG multispectral images

In this article, a segmentation approach for cloud detection in Meteosat Second Generation (MSG) multispectral images is proposed. The proposed algorithm uses recursive segmentation that dynamically reduces the number of classes. This algorithm consists of two steps. First, an initial segmentation of the image is obtained using local fuzzy clustering. The clustering algorithm is formulated by modifying the similarity measure of the standard fuzzy c-means (FCM) algorithm. The new similarity function includes the spectral information as well as the homogeneity and spatial clustering information of each considered pixel. In the second step, a hierarchical region-merging process is used to reduce the number of image clusters. At each iteration, the segmentation algorithm proceeds with a new partition until the final result of the segmentation is obtained. The proposed method has been tested using synthetic and MSG images. It yields a compact and coherent segmentation map, with a satisfactory reproduction of the image contours. Moreover, the different types of clouds are well detected and separated with appropriate accuracy.     

基于图像纹理和小波变换的聚类分割方法

针对纹理图像,本文提出了一种基于图像纹理特征的非学习分割方法。采用小波变换和快速k-means聚类分割算法,减少了整个处理过程的运算量。为了保证分类算法的精确性,运用了总体流量变化最小（Total Variation Flow）[1]非线性去噪方法对图像进行预处理,从而将减小图像噪声污染带来的分割误差。在图像特征的提取上,运用Gabor滤波器原理生成滤波空间,并让图像通过滤波空间而生成特征向量空间。通过制定一个快速寻优策略,从而达到分割图像的目的。     

基于特征散度的自适应FCM图像分割算法

图像分割是模式识别、图像理解、计算机视觉等领域的重要研究内容。基于模糊C均值聚类(FCM)的图像分割是应用较为广泛的方法之一,但其存在距离测度鲁棒性差、需预先给出初始聚类数目、未考虑图像局部相关特性等问题。为克服上述缺点,通过引入特征散度进行距离测度,并结合聚类有效性指数自适应确定初始聚类数目和根据Laws纹理测度提取图像特征等措施,提出了一种新的FCM图像分割算法。实验结果表明,该新算法可以有效地提高图像的分割效果(特别是纹理图像),其分割结果优于现有FCM图像分割方案。     

一种方向Gabor滤波纹理分割算法

结合人眼视觉特性，设计了一种方向Gabor滤波器，该滤波器顾及了纹理图像的方向特性；利用Gabor滤波器的带通技术，抑制次要纹理图像的主频率分量，增强目标纹理图像主频率分量，使滤波输出图像具有较大的类间离散度和较小的类内离散度，将纹理图像的分割转化为传统的图像分割，使图像的分割质量和算法效率都得到了提高。     

What a histogram can really tell the classifier

This paper presents a simple and efficient clustering technique based on the partitioning of the data histogram. The clustering technique was developed in the context of a study of possible unsupervized classification procedures for multispectral earth imagery. The ultimate goal was on-board data compression, the algorithmic production of thematic maps.

The incoming raw spectral data is first reduced to its two principal (Karhunen-Loêve) components, the histogram of which is then partitioned into natural classes on the sole weight of evidence of the global statistics of the imagery. During the course of the study, it became clear that some connection existed between the proposed philosophy and professor Thom's novel theory of “catastrophes”.⁽¹⁾ A simple metric is added to the histogram topology. The metric uses both Shannon's and Fisher's notions of self information. In the domain of definition of the histogram, zones corresponding to the natural classes become separated by a no man's land, an inter-class zone. Under the same formulation, the metric is “Euclidean” on the class zone, “non-Euclidean”, i.e. “Lorentz” on the inter-class zone. This methodology and the underlying philosophy were tested in practice, and encouraging results were obtained.     

Color enhancement of highly correlated images. I. Decorrelation and HSI contrast stretches

Conventional enhancements for the color display of multispectral images are based on independent contrast modifications or “stretches” of three input images. This approach is not effective if the image channels are highly correlated or if the image histograms are strongly bimodal or more complex. Any of several procedures that tend to “stretch” color saturation while leaving hue unchanged may better utilize the full range of colors for the display of image information. Two conceptually different enhancements are discussed: the “decorrelation stretch”, based on principal-component (PC) analysis, and the “stretch” of “hue”-“saturation”-intensity (HSI) transformed data. The PC transformation is scene-dependent, but the HSI transformation is invariant. Examples of images enhanced by conventional linear stretches, decorrelation stretch, and by stretches of HSI transformed data are compared. Schematic variation diagrams or two- and three-dimensional histograms are used to illustrate the “decorrelation stretch” method and the effect of the different enhancements.