CAJ | 학술논문

基于空间邻域信息的二维FCM（2DFCM）算法通过计算像素的邻域平均，将像素的空间位置信息引入分割过程，克服了FCM算法的缺点。但2DFCM算法在计算空间邻域信息时，采用了均值思想，不能有效地体现邻域中其他像素对中心像素的影响程度。为此，本文提出了一种加权的空间邻域信息计算方法，采用模糊聚类的思想，计算邻域中其他像素对中心像素的影响权值，从而得到更加合理的空间邻域信息，进一步提高了FCM的分割性能。实验证明了该方法的有效性。
기우공간린역신식적이유FCM（2DFCM）산법통과계산상소적린역평균，장상소적공간위치신식인입분할과정，극복료FCM산법적결점。단2DFCM산법재계산공간린역신식시，채용료균치사상，불능유효지체현린역중기타상소대중심상소적영향정도。위차，본문제출료일충가권적공간린역신식계산방법，채용모호취류적사상，계산린역중기타상소대중심상소적영향권치，종이득도경가합리적공간린역신식，진일보제고료FCM적분할성능。실험증명료해방법적유효성。
2DFCM algorithm introduces the spatial information into FCM by calculating the mean of neighboring pixels it overcomes the drawbacks of FCM. However 2DFCM algorithm treats each pixel in a neighborhood equally and can not effectively describe the contribution of other pixels to the center pixel. In this paper we propose a new algorithm to calculate the spatial information. The proposed method weights the contribution of each pixel to center pixel in a neighborhood in which the weights is calculated using the idea of fuzzy clustering. The proposed method can obtain more reasonable spatial information and improve the performance of FCM. Experiments demonstrate the effectiveness of the proposed method.