摘要
虚拟人图像是目前研究的重点,其中脑图像尤为突出.在HSV颜色空间中对虚拟人脑图像进行分析,可以加大当前数据与下层数据的区别,以利于目标的分析.给出一种改进的各向异性扩散方程,并构造混合信息场,以降低噪声、过渡区域等因素的影响;使用Ostu算法与一种新颖的快速符号表算法对饱和度信息场与色度信息场进行分类,得到灰质分割结果;并利用解剖学知识、区域信息以及数学形态学知识对亮度场信息进行分析,以修正分割结果,最终将脑组织分离出来.实验结果表明该算法能较精确地得到分割结果.
The Chinese Visible Human Project Research Team from The Third Military Medical University has successfully collected several Chinese visible human（CVH） data sets since October 2002. With these images a learning medicine machine can be developed, but CVH image segmentation is one of the hardest tasks, especially in the brain images. The Chinese visual human brain images are analyzed in HSV color space, which can distinguish different areas in brain clearly. A new fuzzy anisotropic diffusion function is presented, which can diffuse the images while preserving the edge information. With the effect of the fake grey matters, which belong to grey matters located under current image and have similar color with true grey matters, it is hard to get exact grey matters. In order to get exact grey matters from brain images, the Ostu method and a new fast table method are presented to analyze the saturation information, hue information, and value information. Using these methods the grey matters can be segmented exactly. Anatomy inforlnation and region information are fused to analyze the intensity information and confirm the final results correctly. The experiments to segment the Chinese visual human brain images show that the method of this paper can obtain right results in an accuracy way.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2007年第12期2036-2043,共8页
Journal of Computer Research and Development
基金
香港特别行政区政府研究资助局研究基金项目(14185/01E)
香港中文大学研究基金项目(2050345)~~
