• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Conference on image processing >Pattern classification approach to segmentation of digital chest radiographs and chest CT image slices
【24h】

Pattern classification approach to segmentation of digital chest radiographs and chest CT image slices

机译：数字胸片和胸部CT图像切片分割的模式分类方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Abstract: The goal of this research was to develop a segmentationmethod based on a pattern classification approach. Thepattern classification approach consists of classifyingeach pixel into one of several anatomic classes on thebasis of one or more feature values. In this research,three types of locally calculated features are used:gray-level based measures, local difference measuresand local texture measures. A feature selection processis performed to determine which features bestdiscriminate between the anatomic classes. Threeclassifiers are used: a linear discriminant function, ak-nearest neighbor approach and a neural network.Supervised techniques train each classifier to learnthe characterstics of the anatomic classes. Eachclassifier is trained and tested using normal images.The pattern classification approach to imagesegmentation has shown promise for further development.Locally calculated features are important inclassifying pixels, but these alone may not besufficient. A method for incorporating spatialinformation into the classification decision appears toimprove the results and may be necessary for reliablesegmentation. This research also shows that the patternclassification approach may be applied to images fromdifferent modalities. !31
机译：摘要：本研究的目的是开发一种基于模式分类方法的分割方法。模式分类方法包括基于一个或多个特征值将每个像素分类为几种解剖学类别之一。在这项研究中，使用了三种类型的局部计算特征：基于灰度的度量，局部差异度量和局部纹理度量。执行特征选择过程以确定在解剖类别之间最好区分的特征。使用了三个分类器：线性判别函数，近邻ak方法和神经网络。监督技术训练每个分类器以学习解剖学类的特征。每个分类器均使用正常图像进行训练和测试。图像分割的模式分类方法已显示出进一步发展的潜力。局部计算的特征对像素分类很重要，但仅靠这些可能并不足够。一种将空间信息纳入分类决策的方法似乎可以改善结果，对于可靠的细分可能是必要的。这项研究还表明，模式分类方法可以应用于来自不同模态的图像。 ！31

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号