摘要
由于显微成像的精细胞图像对比度较低,背景不均匀,一般阈值分割方法难以得到较好的分割图像。在充分利用图像的空间信息的基础上,结合细胞边缘,提出一种基于边界特征的二维最大熵分割方法,对于目标边缘和目标分别使用不同的阈值分割,给出了算法的步骤和实验结果,实验表明,该算法对于显微精细胞图像分割鲁棒性好,分割准确率高。
Face to the micro - spermatozoa image＇s lower contrast and uneven light of the background, the general threshold segmentation methods are powerless. In order to make full use of the intensity space information, combining with the cell edges, an improved method 2D - Maximum entropy based on the characteristic of spermatozoa boundary is proposed. The different thresholds were used for segmentation. Experiments prove that the method possess good robustness and precision on micro - spermatozoa image.
出处
《计算机与数字工程》
2008年第7期18-20,共3页
Computer & Digital Engineering
关键词
二维最大熵
显微精细胞
图像边缘
图像分割
2D - Maximum entropy, micro - spermatozoa image, image edge, image segmentation
