肺癌细胞学亚型巴氏染色的计算机图像色度学定量分析

用计算机图像处理技术，对痰涂片中巴氏染色的角化性鳞癌细胞（ＫＳＣＣ）、非角化性鳞癌细胞（ＮＫＳＣＣ）和腺癌细胞（ＡＣＣ）胞浆进行了色度学定量分析。测试参数为红（Ｒ）、绿（Ｇ）、蓝（Ｂ）三基包含量和Ｒ、Ｇ、Ｂ的三色系数，同时还测算了胞浆的色度、饱和度、亮度和灰度，并进行了逐步判别分析。结果表明：ＫＳＣＣ的Ｒ、Ｇ、Ｂ及其三色系数、色度、饱和度与ＮＫＳＣＣ和ＡＣＣ相比差异非常显著；三基色的改变较灰度敏感．用三色系数、色度、饱和度对ＫＳＣＣ与ＮＫＳＣＣ、ＫＳＣＣ与ＡＣＣ进行计算机判别，准确度可达９５．３％和９７．１％。这些结果提示：三基色、三色系数及色度、饱和度分析对判别ＫＳＣＣ与ＮＫＳＣＣ、ＫＳＣＣ与ＡＣＣ有重要价值。对有关问题进行了讨论。

QUANTITATIVE CHROMATICS ANALYSES FOR THE COMPUTER IMAGE OF CYTOLOGIC SUBTYPES OF LUNG CANCER STAINED BY THE PAPANICAOLAOU

By means of computer image analysis technique, keratinized squamous carcinoma cell (KSCC), non-keratinized squamous carcinoma cell (NKSCC) and adenocarcinoma cell (ACC)of lung in phlegm smear stained with papanicolaou were analyses and distinguished by quantitative colorimetry of their cytoplasm. The measured parameters were the content of three fundamental colours, Red(R). Green (G). and Blue(B), and the coefficient of these three.The hue. saturation. brightness and grey level were also meased. analyzed and Stepwise discriminant analysis was made. The results showed that the differences of R. G. B and their coeffieients. the hue and saturation between that of KSCC with that of NKSCC and ACC are vern significaton; change of three fundamental colors is mose sensitive than that of grey level.Using computer judgemnt of three fundamental color coefficients, hue and saturation, the corect rate of distinguishing KSCC from NKSCC or KSCC from ACC arched to 95.3% and 97.1% respectively. It is suggestal that analysis of fundamental colors,their coefficients, hue,and satulation is valuble for distinguishing KSCC from NKSCC, or KSCC from ACC. Some issues were discussed.