摘要
为探讨变形雅可比(p=4,q=3)-傅立叶矩对生物小图像的数字化描述性能和医学领域中的应用前景,本实验利用该矩对吞噬不同数目酵母菌的小鼠腹腔巨噬细胞500个变形图像进行特征提取,并用最小平均距离规则对25个比较典型视野图像中选取的300个巨噬细胞显微特征进行自动分割与分类计数实验。结果表明,该不变矩对吞噬不同数目酵母菌的小鼠腹腔巨噬细胞显微特征图像具有很好的数字化描述性能和特征提取能力,平均分类计数准确率达97%。
In order to explore the digitalization properties of Pseudo-Jacobi（p = 4,q = 3）-Fourier moments（PJFM’s） on small bio-images and their practical application in medical fields,the micro characteristics of 500 deformed peritoneal macrophages swallowed different number of yeast fungus were extracted and described by this moment.300 macrophages swallowed different number of yeast fungus from 25 typical visual fields were automatically segmented and classified by using the minimum-mean-distance rule.The results showed that this invariant moment possessed good digitalization and feature extraction abilities for mice peritoneal macrophages swallowed different number of yeast fungus,and the accuracy rate of average count reached 97%.
出处
《中国生物医学工程学报》
CAS
CSCD
北大核心
2010年第2期166-171,共6页
Chinese Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(60967001)
内蒙古自然科学基金(200408020109)
内蒙古自治区高等学校科学研究基金(NJ03037)
关键词
不变矩
腹腔巨噬细胞
图像分割
分类计数
invariant moments
peritoneal macrophages
image segmentation
classification
