| 应用多灰度阈值细胞图象分割法自动识别十类白细胞 |
| |
| 引用本文： | 包含飞,Gelsema ES,den Harink HC,Smeulders AWM. 应用多灰度阈值细胞图象分割法自动识别十类白细胞[J]. 华中科技大学学报(医学版), 1988, 0(3) |
| |
| 作者姓名： | 包含飞 Gelsema ES den Harink HC Smeulders AWM |
| |
| 作者单位： | 同济医科大学基础医学部组织部胚胎学教研室(包含飞)，荷兰自由大学医学信息教研室(Gelsema ES，den Harink HC)，荷兰自由大学医学信息教研室(Smeulders AWM) |
| |
| 摘 要： | 本文报道应用多灰度阈值细胞图象分割法(MSTS)对严格按照随机原则选择的10类白细胞(其中包括5类幼稚细胞)进行自动识别试验的结果。被测细胞总数1149个。如果不计由于研究者本身所造成的误分,那么单纯由计算机造成的误分率仅为6.4％。MSTS摒弃了寻找细胞和细胞核边界这一至今仍不可靠的算法,用自动的细胞图象分割过程替代传统的半自动(对话式)的边界搜索。实验结果表明这一新技术提供了临床快速定量白细胞自动分析的可能性。
|
| 关 键 词： | 血液病 白细胞 计算机 图象处理 模式识别 |
Application of Multiple Sequential Grey Level Thresholds for Segmentation to Automatic Recognition of Ten Types of WBC |
| |
| Bao Hanfei. Application of Multiple Sequential Grey Level Thresholds for Segmentation to Automatic Recognition of Ten Types of WBC[J]. Journal of Huazhong University of Science and Technology(Health Sciences), 1988, 0(3) |
| |
| Authors: | Bao Hanfei |
| |
| Affiliation: | Bao HanfeiDepartment of Histo-Embryology,Faculty of Basic Medical Sciences,Tongji Medical University,WuhanGelsema ES,den Harink HC,Smeulders AWM Department of Medical Informatics,Free University,Amsterdam,Netherlands |
| |
| Abstract: | : The result of a further research on automated classification of 10 types of white blood cells, strictly randomly chosen, through multiple sequential thresholds for segmentation(MSTS), is reported in this paper. The total number of data set is 1149, including 5 types of immature cells. Excluding the misclassifica-tion made perhaps by the investigator himself, only 6.4% of error can be considered definitely due to the computer. The inaccurate method of calculation by searching boundaries cell and its nucleus is thus replaced by MSTS. A new automatic segmentation procedure instead of the traditional semiautomatic one is thereby designed. These improvements may provide a possible technique for clinical use in quick automatic quantitative differention of white blood cells. |
| |
| Keywords: | hematology white blood cell computer image procession pattern recognition |
| 本文献已被 CNKI 等数据库收录！ |
|
