摘要
提出一种基于多图像信息的图像分割方法，利用Ｂａｙｅｓ分类准则，根据每个像素的多图像信息向量将其分为目标或背景，并对动态图像组成的图像序列或静态图像的分割给出了不同的像素分类准则，同时给出了硬件实时实现的原理，该方法适用于图像序列、低对比度图像和其它难于用一种图像信息分割的图像。实验证明了该方法的有效性。
This paper deals with an image segmenting method based on multiple image informa-tion,Depending on the multiple image information of each pixel,a pixel is classified as ob-ject or background with the Bayes criterion. Two separate classifiers are provided for singleimage and dynamic image series,respectively.The real-time hardware design is given,Themethod proposed can be used for the segmentation of image series,low contrast images andother images difficult to segment by only one type of image information;The effectiveness ofthe method has been verified by experiment.
出处
《华中理工大学学报》
CSCD
北大核心
1994年第5期7-10,共4页
Journal of Huazhong University of Science and Technology
关键词
模式识别
图像分割
多图像信息
pattern recognition
image segmentation
Bayes classifieer
multiple image in-formation
image series
