CAJ | 학술논문

基于二维直方图的模糊聚类分割方法,可以有效地抑制噪声的干扰.但是,FCM算法用于图像数据聚类的最大缺陷是运算开销太大,从而限制了该算法在图像分割中的应用.针对这个问题,本文将改进抑制式FCM算法拓展到二维空间,在保持空间信息的基础上,引入了自适应抑制因子来修正不同大小的隶属度,减少了数据的聚类次数,实验结果表明该方法提高了聚类的速度,又有很好的聚类效果.
기우이유직방도적모호취류분할방법,가이유효지억제조성적간우.단시,FCM산법용우도상수거취류적최대결함시운산개소태대,종이한제료해산법재도상분할중적응용.침대저개문제,본문장개진억제식FCM산법탁전도이유공간,재보지공간신식적기출상,인입료자괄응억제인자래수정불동대소적대속도,감소료수거적취류차수,실험결과표명해방법제고료취류적속도,우유흔호적취류효과.
A fuzzy clustering algorithm for image segmentation based on two-dimensional histogram is robust for noise image. Howev-er, the fatal deficiency of FCM algorithm for image data clustering is high operational cost, which limits its application in image seg-mentation. To solve this problem, in the paper the improved suppressed FCM is extended to the two-dimensional space and the de-gree of membership is amended by adaptive inhibitory factor, which induce the number of data clustering while maintaining spatial information. Experimental results indicate that the new algorithm increases the speed of clustering and has a good clustering effect.