摘要
研究基于图论的最短路径算法与加权直方图方法,结合快速模糊聚类思想,提出一种改进的快速模糊C-均值(FCM)图像分割算法。利用最短路径算法对图像进行初始化处理,使处理后的灰度值准确反映邻域像素对原像素的影响。通过加权直方图改变灰度变化剧烈区域像元在图像分割中的影响程度,并自适应寻找初始聚类中心。实验结果表明,该算法能快速准确地分割图像,具有较强的抗噪性。
This paper researches the shortest path algorithm and the weighted histogram image segmentation method. Combined with rapid fuzzy clustering thought, it presents an improved rapidly Fuzzy C-means（FCM） image segmentation algorithm. Through the shortest path algorithm, it initializes the image to make it accurately reflect the pixel neighborhood. It changes a weighted histogram to influence degree of gray strong regional pixel in the image segmentation and identifies a clustering center by adaptive weighted histogram. Experimental result shows that this algorithm can rapidly and accurately segments images and has strong antinoise.
出处
《计算机工程》
CAS
CSCD
2012年第8期192-194,197,共4页
Computer Engineering
基金
中央高校基本科研业务费基金资助项目(CDJXS11100032)
关键词
最短路径
加权直方图
模糊聚类
邻域信息
邻域像素
抗噪性
the shortest path
weighted histogram
fuzzy clustering
neighborhood information
neighborhood pixel
antinoise
