Fuzziness parameter selection in fuzzy c-means: The perspective of cluster validation

Fuzzy c-means （FCM） algorithm is an important clustering method in pattern recognition, while the fuzziness parameter, m, in FCM algorithm is a key parameter that can significantly affect the result of clustering. Cluster validity index （CVI） is a kind of criterion function to validate the clustering results, thereby determining the optimal cluster number of a data set. From the perspective of cluster validation, we propose a novel method to select the optimal value of m in FCM, and four well-known CVIs, namely XB, VK, VT, and SC, for fuzzy clustering are used. In this method, the optimal value of m is determined when CVIs reach their minimum values. Experimental results on four synthetic data sets and four real data sets have demonstrated that the range of m is [2, 3.5] and the optimal interval is [2.5, 3].     

舌像颜色校正算法对比研究

The color of tongue coating is an important feature in tongue diagnosis of traditional Chinese Medicine. In order to solve the problem of color distortion in image acquisition of TCM tongue image in natural environment, when shooting the tongue image, the standard color card is placed beside the tongue to shoot together, and the polynomial regression algorithm is used to correct the color of the image. Experimental results show that the proposed method is more accurate than perfect reflection algorithm, gray world algorithm, color deviation detection and color correction algorithm.     

Clustering by data competition

Clustering analysis is an unsupervised method to find out hidden structures in datasets.Most partitional clustering algorithms are sensitive to the selection of initial exemplars,the outliers and noise.In this paper,a novel technique called data competition algorithm is proposed to solve the problems.First the concept of aggregation field model is defined to describe the partitional clustering problem.Next,the exemplars are identified according to the data competition.Then,the members will be assigned to the suitable clusters.Data competition algorithm is able to avoid poor solutions caused by unlucky initializations,outliers and noise,and can be used to detect the coexpression gene,cluster the image,diagnose the disease,distinguish the variety,etc.The provided experimental results validate the feasibility and effectiveness of the proposed schemes and show that the proposed approach of data competition algorithm is simple,stable and efficient.The experimental results also show that the proposed approach of data competition clustering outperforms three of the most well known clustering algorithms K-means clustering,affinity propagation clustering,hierarchical clustering.     

Hierarchical adaptive stereo matching algorithm for obstacle detection with dynamic programming

An adaptive weighted stereo matching algorithm with multilevel and bidirectional dynamic programming based on ground control points （GCPs） is presented. To decrease time complexity without losing matching precision, using a multilevel search scheme, the coarse matching is processed in typical disparity space image, while the fine matching is processed in disparity-offset space image. In the upper level, GCPs are obtained by enhanced volumetric iterative algorithm enforcing the mutual constraint and the threshold constraint. Under the supervision of the highly reliable GCPs, bidirectional dynamic programming framework is employed to solve the inconsistency in the optimization path. In the lower level, to reduce running time, disparity-offset space is proposed to efficiently achieve the dense disparity image. In addition, an adaptive dual support-weight strategy is presented to aggregate matching cost, which considers photometric and geometric information. Further, post-processing algorithm can ameliorate disparity results in areas with depth discontinuities and related by occlusions using dual threshold algorithm, where missing stereo information is substituted from surrounding regions. To demonstrate the effectiveness of the algorithm, we present the two groups of experimental results for four widely used standard stereo data sets, including discussion on performance and comparison with other methods, which show that the algorithm has not only a fast speed, but also significantly improves the efficiency of holistic optimization.     

Relaxation algorithm of piecing-error for sub-images

During the process of automatic image recognition or automatic reverse design of IC,people often encounter the problem that some sub-images must be pieced together into a whole image,In the traditional piecing algorithm for subimages,a large accumulated error will be made.In this paper,a relaxation algorithm of piecing -error for sub-images is presented.It can eliminate the accumulated error in the traditional algorithm and greatly improve the quality of pieced image.Based on an initial pieced image,one can continuously adjust the center of every sub-image and its angle to lessen the error between the adjacent sub-images,so the quality of pieced image can be improved.The presented results indicate that the proposed algorithm can dramatically decrease the error while the quality of ultimate pieced image is still acceptable.The time complexity of this algorithm is O(nlnn).     

An accelerated K-means clustering algorithm using selection and erasure rules

The K-means method is a well-known clustering algorithm with an extensive range of applications,such as biological classification,disease analysis,data mining,and image compression.However,the plain K-means method is not fast when the number of clusters or the number of data points becomes large.A modified K-means algorithm was presented by Fahim et al.(2006).The modified algorithm produced clusters whose mean square error was very similar to that of the plain K-means,but the execution time was shorter.In this study,we try to further increase its speed.There are two rules in our method:a selection rule,used to acquire a good candidate as the initial center to be checked,and an erasure rule,used to delete one or many unqualified centers each time a specified condition is satisfied.Our clustering results are identical to those of Fahim et al.(2006).However,our method further cuts computation time when the number of clusters increases.The mathematical reasoning used in our design is included.     

A new approach to obtain K-means initial clustering center based on fuzzy granular computing

The traditional K-means is very sensitive to initial clustering centers and the clustering result will wave follow the different initial input. To remove this sensitivity, a new method is proposed to get initial clustering centers. This method is as follows： provide a normalized distance function d（di,dj） in the fuzzy granularity space of data objects, then use the function to do a initial clustering work to these data objects who has a less distance than granularity dλ, then get the initial clustering centers. Approved by the test, this method has such advantages on increasing the rate of accuracy and reducing the program times.     

障碍空间中基于网格的不确定数据聚类算法

为了解决已有研究成果无法有效解决动态障碍空间中的不确定数据聚类问题,根据障碍集合是否发生变化,分别解决静态障碍和动态障碍空间下的聚类问题。提出了静态障碍空间中的不确定数据聚类算法(DBSCAN clustering algorithm for static obstacles in grid space,STA_GOBSCAN)、障碍物动态增加情况下的不确定数据聚类算法(DBSCAN clustering algorithm for dynamic increase of obstacles in grid space,DYN_GOCBSCAN)、障碍物动态减少情况下的不确定数据聚类算法(DBSCAN clustering algorithm for dynamicreduction of obstacles in grid space,DYN_GORBSCAN)和障碍物动态移动情况下的不确定数据聚类算法(DBSCAN clustering algorithm for dynamic movement of obstacles in grid space,DYN_GOMBSCAN),采用KL距离对不确定数据进行相似性度量,并利用网格对数据空间进行划分。理论研究和实验结果表明所提出的算法具有较高的效率和准确率。     

Vector morphological operators in HSV color space

In HSV color space,the current vector morphological operators have low capability to reduce color noise caused by hue and saturation in color image processing.Because they sort the color pixels according to the hierarchical ordering of V,S,H,which is against the equal principle of the three channels in color image processing.A novel vector ordering based on the combination of H,S and V is proposed in this paper,and the associated vector morphological erosion,dilation and composite filtering operators are defined.Compared with the popular vector morphological operators,experimental results show that the new operators can reduce the color noise effectively without any new color pixels while preserving the image details.And the filtered images have higher peak signal-to-noise ratio(PSNR) and lower mean absolute error(MSE).     

自适应FCM算法在图像分割中的应用研究

针对目前还没有较好的方法确定模糊C均值FCM聚类中C值和各个初始聚类中心这一问题,提出一种先用进化聚类快速确定初始聚类中心和聚类个数C,后用模糊C均值FCM聚类的算法,算法时间复杂度和空间复杂度与C均值FCM基本相当。应用该算法在人物图像和遥感图像中进行了分割实验验证,算法在分割的准确性和模糊边界的分隔上取得令人满意的效果。     

基于分裂式K均值聚类的图像分割方法

模糊C均值聚类(FCM)算法是一种有效的无监督图像分割方法,适用于任意分类数,不需要预知图像特征,但其聚类效果直接受待分类样本噪声和分类初始条件的影响。因此,提出了一种适用于彩色图像分割的分裂式K均值聚类(FKM)算法,该算法首先使用中值滤波对分类样本去噪,然后使用一种分裂聚类法对图像样本进行预分类,得到一组样本集初始划分,最后以这组划分为起点,使用基于概率距离的K均值聚类对图像分割进行迭代优化。实验结果表明,该算法可以避免FCM的误分类,诸如陷于中心死区、中心重叠和局部极小值,而且提高了分割速度。     

基于蚁群和带空间约束FCM的荔枝图像分割算法

准确地提取荔枝果实的完整轮廓对采摘机器人自动识别与采摘至关重要。以蚁群和模糊C均值（FCM）聚类为理论基础,选用符合荔枝颜色特性的L*a*b*颜色空间,提出一种基于蚁群和带空间约束FCM的荔枝图像分割算法。该算法利用L*a*b*颜色空间的a*通道正轴代表红色和负轴代表绿颜色进行初始分割,然后利用蚁群聚类算法全局性和鲁棒性的优点确定FCM的聚类中心,用引入空间约束的FCM完整地分割出荔枝果实。实验结果表明此方法实现了荔枝图像完整地分割,并且满足了采摘机器人后续的荔枝识别与采摘,对成熟荔枝分割的正确率达到了87%。     

基于模糊C均值聚类的多分量彩色图像分割算法

以模糊C均值(FCM)聚类理论为基础,选用符合人眼视觉特性的HSI颜色空间,提出了一种新的多分量彩色图像分割算法。该算法首先结合数据分布特点确定出H分量与I分量的初始聚类中心;然后利用FCM聚类技术对H分量、I分量进行分类处理,以得到不同分量的像素点隶属度;最后,将所得到的不同分量像素点隶属度组织成2维特征,并以此进行模糊聚类图像分割。实验结果表明,该算法可有效提高图像分割效果,其分割结果优于传统FCM聚类图像分割方案。     

空间模糊C均值聚类的神经切片图像分割方法

周围神经切片显微图像具有背景复杂、区域不连续和光照不均匀等特点,应用经典的图像分割算法难以取得有效的分割结果。通过结合初始隶属度概率函数和空间距离来设计空间函数而得到的SFCM聚类算法,并提出SFCM彩色图像分割方法。把图像从RGB颜色空间转换到HIS颜色空间。采用聚类有效性分析指标在直方图快速FCM算法中为HSI各分量确定分类数目和获取SFCM初始化参数。对HIS各分量分别进行SFCM聚类,合并各分量并转换回RGB彩色空间以显示结果。实验结果表明,与标准FCM聚类分割算法相比,新方法能更有效地分割区域不连续的神经切片显微图像。     

基于改进蝙蝠算法优化的FCM聚类算法

针对传统模糊C-均值（Fuzzy C-Means, FCM）聚类算法隐含假设各个样本和各维属性对聚类结果作用相同,导致算法聚类性能降低,以及对初始中心点敏感且易陷入局部最优的问题,提出一种基于改进蝙蝠算法优化的FCM聚类算法。该算法首先采用混沌映射和速度权重来改进蝙蝠算法,然后利用改进蝙蝠算法确定FCM算法的初始聚类中心,最后根据各个样本和各维属性对聚类结果作用不同,采用样本和属性加权法对FCM算法的目标函数重新设计。实验结果表明,改进算法表现出较好的聚类效果。     

一种用于彩色图象量化的快速聚类算法

为了对彩色图象进行有效地压缩处理,提出了一种基于模式识别技术的图象量化新算法（FSCAMMD）,该算法首先把彩色图象中的颜色样本归为一类,并采用最大频度与类内最小距离最大相结合的方法选取初始类代表点--初始值优选法;然后采用欧氏距离聚类准则及重心法,求得新聚类域中心的向量值,从而得到了令人满意的量化效果。该算法不仅克服了SCA算法对聚类中心初始值选取的不足,较大幅度地减少了彩色图象量化后的总方差以及颜色失真度,而且较好地解决了重建彩色图象的整体层次与局部细节之间的矛盾,其量化效果优于SCA和其他一些聚类量化算法。     

基于GFCMA算法的医学彩色图像分割

医学图像分割是医学图像分析的关键步骤,经典的模糊C-均值聚类算法(FCM)是常用方法,但其依赖于初始聚类中心的选择,通常存在局部收敛的缺陷。通过与遗传算法(GA)结合而成的遗传模糊C-均值聚类算法(GFCMA),采用RGB颜色空间,能够得到全局最优解,并在此基础上实现了医学彩色图像分割和特定目标提取,取得良好分割效果。     

基于HSI空间模糊聚类的彩色图像分割新方法

模糊C均值(FCM)被广泛应用于彩色图像分割中,但传统的模糊C均值由于没有考虑空间信息,因此对噪声特别敏感。针对此问题,提出了一种在HIS颜色空间结合像素邻域空间信息的模糊聚类新方法。实验结果表明,此方法对高噪声图像有较好的处理结果。     

基于快速二维熵的加权模糊C均值聚类图像分割

提出了一种结合快速二维熵和加权模糊C均值聚类的图像分割方法。采用快速二维熵算法对实际图像进行初步分割求得目标和背景的中心,然后采用样本点像素与其邻域灰度像素的差别表征该样本点对分类的影响程度,最后利用加权模糊C均值聚类算法完成图像分割。该方法一方面解决了传统的模糊C均值聚类算法对初始值敏感的问题,另一方面克服了传统的聚类算法对数据集进行等划分的缺陷。实验结果表明,该方法不仅具有良好的收敛性,而且还可以有效地把目标从背景中分割出来,具有重要的实际应用价值。