微型机与应用
微型機與應用
미형궤여응용
MICROCOMPUTER & ITS APPLICATIONS
2012年
20期
33-35
,共3页
图像分割%模糊C均值%空间信息
圖像分割%模糊C均值%空間信息
도상분할%모호C균치%공간신식
image segmentation%fuzzy C-means%spatial information
针对传统的模糊C均值聚类算法（FCM）在图像分割中对噪声十分敏感这一局限性，提出一种自适应的FCM图像分割方法。该方法充分考虑图像像素的灰度信息和空间信息，根据像素的空间位置自适应地计算一个合适的相似度距离来进行聚类分割图像。实验结果表明，与传统的FCM相比，该方法能显著提高分割质量，尤其是能提高对于图像噪声的鲁棒性和分割图像区域边缘的准确性。
針對傳統的模糊C均值聚類算法（FCM）在圖像分割中對譟聲十分敏感這一跼限性，提齣一種自適應的FCM圖像分割方法。該方法充分攷慮圖像像素的灰度信息和空間信息，根據像素的空間位置自適應地計算一箇閤適的相似度距離來進行聚類分割圖像。實驗結果錶明，與傳統的FCM相比，該方法能顯著提高分割質量，尤其是能提高對于圖像譟聲的魯棒性和分割圖像區域邊緣的準確性。
침대전통적모호C균치취류산법（FCM）재도상분할중대조성십분민감저일국한성，제출일충자괄응적FCM도상분할방법。해방법충분고필도상상소적회도신식화공간신식，근거상소적공간위치자괄응지계산일개합괄적상사도거리래진행취류분할도상。실험결과표명，여전통적FCM상비，해방법능현저제고분할질량，우기시능제고대우도상조성적로봉성화분할도상구역변연적준학성。
Traditional fuzzy C-means clustering algorithm(FCM) segmentation of the image is very sensitive to the noise. In or- der to overcome this limitation, an adaptive FCM image segmentation is proposed. This algorithm takes full account of the image pixel gray-scale and spatial information, adaptively calculates a suitable distance based on the location of the pixel to the clustering image segmentation. The experimental resuhs show that compared with traditional FCM, the proposed method can significantly im- prove the segmentation quality, in particular to improve the robustness to the image noise and the accuracy of the image area edge of split.
