Suppr超能文献

定量核图像分析:石蜡切片中正常、增生性和恶性子宫腺的鉴别。IV. 判别分析中马尔可夫链纹理特征的应用。

Quantitative nuclear image analysis: differentiation between normal, hyperplastic, and malignant appearing uterine glands in a paraffin section. IV. The use of Markov chain texture features in discriminant analysis.

作者信息

Diegenbach P C, Baak J P

出版信息

Eur J Obstet Gynecol Reprod Biol. 1978 Jun;8(3):157-62. doi: 10.1016/0028-2243(78)90067-9.

Abstract

Discriminant analysis was applied to Markov chain texture features and elementary features calculated from data from microscope photometry of nuclei in a paraffin section. The results from measurements on the nuclei of morphologically normal, atypical hyperplastic and carcinomatous uterine glands were used in discriminant analysis. With this technique it is possible to classify up to 88.1% of the nuclei correctly in one of the three groups of uterine glands. The discriminating power of several smaller subsets indicated that with more than 28 features there is hardly any increase in discriminating power. Discriminant analysis with a selection from elementary and Markov chain features provides objective criteria of assistance in histopathological diagnosis.

摘要

判别分析应用于马尔可夫链纹理特征和从石蜡切片中细胞核的显微镜光度测量数据计算得到的基本特征。对形态学上正常、非典型增生和癌性子宫腺细胞核的测量结果用于判别分析。使用该技术可将高达88.1%的细胞核正确分类到三组子宫腺中的一组。几个较小子集的判别能力表明,超过28个特征时,判别能力几乎没有增加。从基本特征和马尔可夫链特征中进行选择的判别分析为组织病理学诊断提供了客观的辅助标准。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验