基于特征参数和云模型相似性的医学图像自动归档-《中国医学物理学杂志》

文章信息/Info

Title:: Automatic archiving of medical images based on feature parameters and cloud model similarity

Author(s):: WU Di1; HU Sheng2; 1. School of Basic Medical Science, Shaanxi University of Chinese Medicine, Xianyang 712046, China 2. School of Mechanical and Electrical Engineering, Xi′an Polytechnic University, Xi′an 710048, China

Keywords:: Keywords: medical image image classification and archiving feature parameter gray level coocurrence matrix cloud model similarity degree

摘要:: 针对医学图像由于特征对比度不高导致医学影像归档不准确的问题，提出一种基于特征参数和云模型相似性的医学影像图像自动归档方法。首先，通过提取医学图像特征参数并进行增强处理获取不同影像图像的特征参数；然后，利用特征图像生成云滴特征集合，构建能够反映医学影像特征的三元组图像云模型，在此基础上定义云模型相似度并通过计算待归档影像图像对各病理特征图像的相似度大小，实现医学图像的自动识别和归档。通过实验进行验证，结果显示所提方法可以有效地对医学影像图像进行自动分类，为医学影像数据的自动分类归档提供新思路。

Abstract:: Abstract: An automatic archiving method for medical images is proposed based on feature parameters and cloud model similarity to address the problem of inaccurate image archiving due to low feature contrast. The feature parameters of different images are obtained by extraction of the feature parameters from medical images and image enhancement. Then, the feature images are used to generate cloud droplet feature sets for constructing a ternary image cloud model that can reflect the image characteristics. The cloud model similarity degree is defined, and the similarity of various images to be archived is calculated to realize the automatic classification and archiving of medical images. The experiment confirms that the proposed method can effectively classify medical images automatically, which provide new ideas for the automatic classification and archiving of medical imaging data.