摘要
〔目的〕研究并开发疟原虫图像自动识别软件系统,以满足国境口岸疟疾疫情的检疫需求。〔方法〕首先,建立疟原虫图像标准模板库;然后,图像经消噪处理,利用分水岭算法进行图像分割,统计出细胞个数;最后,系统对染色图像特征学习,利用已获得的染色图像特征点进行染色图像提取,从而报告出阳性感染疟原虫细胞。〔结果〕对疟疾薄血膜片总体识别率达90%。〔结论〕疟原虫图像自动识别软件系统的建立,有助于提高国境口岸疟疾诊断的效率与准确率,为疟疾数字化诊断奠定了基础。
Objective To research and develop automatic image identifying technology system of parasite malarial for quarantine at border port. Method A standarized plasmodium template storehouse was established, and the image chirp eliminated and divided them by using watershed algorithm.Finally, the characteristic image was extracted after learning its features, and thus the malarial positive cell was reported. Result The total recognition rate was 90%. Conclusion The stablishment of automatic image identifying technology system of plasmodium can improve efficiency and accuracy for diagnosis of malarial, and create a method of digital diagnosis for malarial.
出处
《中国国境卫生检疫杂志》
CAS
2008年第4期241-244,共4页
Chinese Journal of Frontier Health and Quarantine
基金
江苏出入境检验检疫局科研基金项目(2005KJ26)
关键词
疟原虫
图像
自动识别
Plasmodium
Image
Automatic identifying technology
