液基细胞学图文系统的设计与实现  

Research on Thin-layer Cytology Assistant Screening System

作　　者：陆尧胜[1] 梁嘉铄[1] 宋立国 

机构地区：[1]暨南大学信息科技学院,广东广州510632 [2]三瑞医疗器械有限公司,广东广州510520

出　　处：《中国医学物理学杂志》2009年第6期1559-1564,共6页Chinese Journal of Medical Physics

摘　　要：目的:宫颈涂片筛检异常细胞技术为降低宫颈癌患者的死亡率起了重要作用。然而,一直以来涂片经肉眼在显微镜下检查,结果存在较高的假阳性率和假阴性率。方法:为了解决这一问题,研制了液基细胞图文系统,硬件方面对传统的显微图像分析系统进行了改进;软件方面引入了灰度变换、边缘检测、图像拼接等图像处理方法,利用神经网络技术实现了液基细胞图像的采集、拼接、处理、测量和识别等功能。结果:使用本系统对150例TCT宫颈细胞样品检测,与病理专家的人工分析结果进行比较,结果相符率超过90%。结论:本系统用于辅助病理学医生进行传统巴氏涂片和液基细胞学涂片的检查,可大大提高宫颈涂片检查的准确率,降低医生的劳动强度、提高工作效率。Objective： Cervical smear screening for abnormal cells plays an important role in reducing the mortality of cervical cancer. However, all smears must be approved by the naked eyes using microscopes, and there has been a higher false-positive rate and false-negative rate. Methods： To solve this problem, the thin-layer cytology assistant screening system was developed. On the hardware side, the traditional computer-aided analysis system was improved; and on the software side, gray level transformation, edge detection, image mosaic, and some other image processing methods were introduced; and by the use of neural net- work, the system could realize the function of acquisition, mosaic, process, measure and recognition of the liquid-based cytology image. Results： This system was brought into the test of 150 thin-layer cytology technology（TCT） cervical cell samples, and the agreement between the results of computer-aided analysis and the manual analysis by pathology experts exceeded 90%. Conclusion： Both the traditional Pap smear and TCT preparation smear can be diagnosed by this system. As a result, the false-positive rate and false-negative rate of the cervical smear screening will be significantly reduced, meanwhile, work efficiency of pathologist can be greatly improved.

关 键 词：宫颈涂片 宫颈癌 巴氏涂片 液基细胞学 神经网络 

分 类 号：R318.04[医药卫生—生物医学工程] TP391[医药卫生—基础医学]