基于图论的Normalized Cut图像分割方法研究

基于图论的图像分割是近年来国际上研究图像分割算法领域中一个新的热点研究方向,Normalized Cut（规一化分割）是基于图论的分割方法中比较典型的一种。本文介绍Normalized Cut方法的相关内容,分析它的优缺点并提出几种有效的改进方法。最后,实验显示Normalized Cut在特定图片的分割上取得了不错的效果。     

多分辨率下基于Normalized Cut的图像分割

提出了一种图像分割的新方法。该方法首先对低分辨率下的小图像使用Normalized Cut,快速地提取这幅小图像中高层视觉信息,然后将这一高层信息有效地还原回高分辨率下并进行平滑处理。实验中得到了很好的全局分割效果。该方法既充分利用了Normalized Cut的全局分割特性,又有效地克服了Normalized Cut的大运算量问题,提高了图像分割的速度。     

基于分水岭算法和图论的图像分割

提出一种新的基于分水岭算法和图论的图像分割方法FWTN(First Watershed Then Normalized cut),以克服传统的分水岭算法造成的图像过度分割。FWTN方法在分水岭算法之后,用Normalized Cut方法在区域之间进行分割,产生最终所分割的图像。FWTN成功地解决了过分割问题,同时还对Normalized Cut方法中的NP-hard问题有所改进。实验结果表明该算法是一种切实可行的图像分割方法。     

融合Ncut方法的脊椎图像OTSU分割算法

提出一种融合Normalized Cut方法和OTSU方法的脊椎图像分割算法。首先通过Normalized Cut方法得到粗略的分割结果,然后根据脊椎的形状特征提取出有效的脊椎区域,在有效的脊椎区域进行OTSU分割算法,并通过边缘检测算法得到脊椎内部的轮廓边界。实验结果证明该方法可以较为精确地检测出椎骨轮廓的边界,有效地消除了伪边缘的影响。提出了按需分割的概念,首先消除背景区域的影响,然后在有效区域进行更为精细的处理,这样可以较大程度提高算法的处理速度。     

基于四叉树与图割的遥感图像分割方法

针对图割算法实时性差的问题,提出一种基于四叉树与图割的遥感图像分割方法。用四叉树分割方法将遥感图像分割成多个区域,根据区域的邻接性合并相似小区域,结合区域间的灰度和位置信息,利用Normalized Cut方法在区域间进行划分,完成图像的分割。实验结果证明该方法能有效消除四叉树方法的过分割现象。     

基于Normalized Cut的图像分割改进算法

用Normalized Cut（N-Cut）准则分割图片时,会出现诸如过分割或者欠分割的不理想情况。在N-Cut结果的基础上提出了一种改进算法：利用合并-分裂方法,将颜色和纹理相似的区域合并,并对某些区域做进一步N-Cut分割。实验证明,改进后的算法不仅保留了N-Cut方法的优点,而且从一定程度上解决了过分割和欠分割的缺点,能够处理一些自然界的彩色图片,得到了比较理想的分割结果。     

基于人类视觉系统区域分裂的图割遥感图像分割算法研究

针对传统Normalized Cut(NC)在分割图像过程中实时性差的特点,提出一种使用人类视觉系统(HVS)区域分裂的图割方法.根据HVS特性用区域分裂算法将遥感图像分割成多个小区域,再结合各个小区域的灰度和空间信息从全局角度用NC方法在区域间进行划分,完成图像的最终分割.实验表明,相对于传统的NC方法,该算法在获得...     

Ncut在颅脑MRI肿瘤提取中的应用研究

颅脑肿瘤自身边缘包含重要的病变信息,提取脑肿瘤区域,对脑部疾病的诊断和治疗具有重要意义。Ncut(Normalized Cut)是基于图理的典型分割方法。将Ncut算法应用到颅脑MRI肿瘤图像的分割中,针对不同颅脑MRI肿瘤图像,进行相关参数测试,选择合适的权重及参数,进行颅脑MR肿瘤的提取。通过利用Matlab进行仿真测试可知Ncut方法能够提取出肿瘤所在的基本轮廓,取得较好效果。     

基于结构张量的GrabCut图像分割算法

传统的Grab Cut图像分割方法大多基于图像的像素值建立图模型,未考虑到彩色图像中丰富的纹理信息。为此,提出一种新的Grab Cut模型图像分割算法。对比基于结构张量的Grab Cut分割方法和传统Grab Cut分割方法的结果,利用结构张量和像素值构建紧缩的结构张量。为提高计算的简洁性和高效性,将Grab Cut方法构建的混合高斯模型扩展到张量空间,并用Kullback-Leible散度代替常用的黎曼度量。在合成纹理图像和自然图像上进行的实验结果表明,与Carsten Rother,GACWRF等方法相比,该算法具有更精确的分割效果,不仅实现了纹理信息与颜色信息的无参融合,而且提高了计算效率。     

基于超像素的人工神经网络图像分类

基于人工神经网络对图像标签分类,为简化后续数据处理,先用Normalized Cut将图像分割为超像素,提取特征向量,通过输入训练样本集,对网络进行训练,在最小均方误差意义下得到网络参数,最后在Matlab的仿真实验中基于不同隐藏层节点,使用BP神经网络模型对图像超像素进行分类。     

一种融合区域生长与图论的图像分割方法

该文提出一种融合区域生长与图论的图像分割方法,一般的基于区域的分割方法在区域生长完成之后需要进行区域的合并,以消除过分割现象。该文的方法在区域生长完成之后,用NormalizedCut方法在区域之间进行分割,产生最终所分割的图像。在方法上区域生长方法考虑的是图像的局部信息,NormalizedCut方法考虑的是图像的全局信息,该文的方法融合了两者的优点。该文的算法主要以灰度图像为研究对象,实验结果表明可以取得很好的分割效果。     

Normalized Cuts Revisited: A Reformulation for Segmentation with Linear Grouping Constraints

Indisputably Normalized Cuts is one of the most popular segmentation algorithms in pattern recognition and computer vision. It has been applied to a wide range of segmentation tasks with great success. A number of extensions to this approach have also been proposed, including ones that can deal with multiple classes or that can incorporate a priori information in the form of grouping constraints. However, what is common for all these methods is that they are noticeably limited in the type of constraints that can be incorporated and can only address segmentation problems on a very specific form. In this paper, we present a reformulation of Normalized Cut segmentation that in a unified way can handle linear equality constraints for an arbitrary number of classes. This is done by restating the problem and showing how linear constraints can be enforced exactly in the optimization scheme through duality. This allows us to add group priors, for example, that certain pixels should belong to a given class. In addition, it provides a principled way to perform multi-class segmentation for tasks like interactive segmentation. The method has been tested on real data showing good performance and improvements compared to standard normalized cuts.     

Mean shift spectral clustering

In recent years there has been a growing interest in clustering methods stemming from the spectral decomposition of the data affinity matrix, which are shown to present good results on a wide variety of situations. However, a complete theoretical understanding of these methods in terms of data distributions is not yet well understood. In this paper, we propose a spectral clustering based mode merging method for mean shift as a theoretically well-founded approach that enables a probabilistic interpretation of affinity based clustering through kernel density estimation. This connection also allows principled kernel optimization and enables the use of anisotropic variable-size kernels to match local data structures. We demonstrate the proposed algorithm's performance on image segmentation applications and compare its clustering results with the well-known Mean Shift and Normalized Cut algorithms.     

Investigations on adaptive connectivity and shape prior based fuzzy graph-cut colour image segmentation

Image segmentation is a challenging problem in computer vision with wide application. It is a process which considers the similarity criterion required to separate an image into different homogenous connected regions. First, an Optimized Adaptive Connectivity and Shape Prior in Modified Graph Cut Segmentation method has been applied to handle the structural irregularities in images. Second, an Optimized Adaptive Connectivity and Shape Prior in Modified Fuzzy Graph Cut Segmentation (Opac-MFGseg) is proposed to partition the images based on feature values. In this method, a fuzzy rule based system is used with optimization algorithm to provide the information on how much a specific feature is involved in image boundaries. The graph obtained from this fuzzy approach is further used in adaptive shape prior in modified graph cuts framework. Moreover, this method supports moving images (videos). In such a situation, a fully dynamic method called Optimized Adaptive Connectivity and Shape Prior in Dynamic Fuzzy Graph Cut Segmentation (Opac-DFGseg) method is proposed for the image segmentation. The effectiveness of the Opac-MFGseg and Opac-DFGseg methods is tested in terms of average sensitivity, precision, area overlap measure, relative error, and accuracy and computation time.     

结合形状约束的Graph Cut行人分割

传统的Graph Cut算法没有对目标的形状予以限制,很难得到语义化的分割结果,即无法保证分割出来的是“行人”。针对该问题提出一种结合形状和底层特征的Graph Cut算法。对于行人分割,用大量真实行人轮廓来表达“行人”的先验形状,对Graph Cut分割算法予以约束,同时构建一个行人模板的层次树以减少匹配时间;并且提出一种区分性的外观模型来替换原来的外观模型。实验结果证明,该算法的分割结果明显优于传统Graph Cut算法的分割结果,所得到的轮廓与真实的行人轮廓比较吻合。     

一种改进的One-Cut交互式图像分割算法

GrabCut算法作为一种典型的交互式彩色图像分割算法,是计算机图像领域中的重要技术手段。然而随着大数据时代的到来,图像数据种类和数量都呈指数级增长,显著地增加了图像分割的任务量,对图像分割效率提出了更高的要求。针对GrabCut算法图像分割效率及精度低的问题,提出了一种改进的One Cut交互式图像分割算法。首先采用One Cut的L1距离项构建能量函数避免GrabCut算法所面临的NP hard问题。然后改进能量函数中表观重叠惩罚项,并结合颜色直方图加速技术,优化网络图结构,显著降低网络图的复杂度,从而提高图像分割的效率及精度。实验结果表明,改进后的One Cut图像分割算法显著提升了图像分割效率,提高了分割精度,得到了较好分割结果。