基于形状纹理特征的食管癌和肝包虫病图像分类

目的:探讨用K最近邻(KNN)分类算法对食管癌X射线图像和肝包虫CT图像的Hu不变矩形状特征和小波变换纹理特征进行分类研究。方法:利用Hu不变矩算法和小波变换算法对食管癌X射线图像和肝包虫CT图像提取特征,用KNN分类器对特征值进行分类以验证所提取特征的分类能力。结果:对于食管癌X射线图像使用Hu不变矩算法提取形状特征具有较好的分类性能,对于肝包虫CT图像使用小波变换算法提取纹理特征具有较好的分类性能。结论:Hu不变矩形状特征结合KNN分类器的研究方法为新疆哈萨克族食管癌的分型提供一定的依据,小波变换纹理特征结合KNN分类器的研究方法为地方性肝包虫病的分型提供一定的依据,同时为计算机辅助诊断系统的研发奠定基础。

Image classification of esophageal cancer and hepatic hydatid disease based on shape and texture features

Abstract: Objective To explore the classification of Hu moment invariant features and wavelet transform texture features of the X-ray image of esophageal cancer and the CT image of hepatic hydatid disease by K nearest neighbor (KNN) classification algorithm. Methods Hu moment invariant and wavelet transform algorithms were used to extract the features of the X-ray image of esophageal cancer and the CT image of hepatic hydatid disease. Moreover, KNN classifier was used to classify the feature values for verifying the classification ability of the extracted features. Results For the X-ray image of esophageal cancer, Hu moment invariant algorithm had good classification performance in extracting shape features. Using wavelet transform algorithm to extract texture features of the CT image of hepatic hydatid disease also had preferable classification performance. Conclusion Hu moment invariant features combined with KNN classifiers provide a basis for the classification of esophageal cancer in Xinjiang Kazakh; and wavelet transform texture features combined with KNN classifiers provide a basis for the classification of endemic hepatic hydatid disease. The study also lays the foundation for the development of computer-aided diagnosis system.