【作者】 彭莎； 吴效明； 杨荣骞；

【Author】 Peng Sha Wu Xiao-ming Yang Rong-qian (School of Bioscience & Bioengineering, South China University of Technology, Guangzhou 510006, Guangdong, China)

【机构】 华南理工大学生物科学与工程学院生物医学工程系；

【摘要】 数字抠图技术解决了单个像素点上前景和背景色彩分离的问题,在应用上具有广阔的前景。本文针对数字抠图技术中传统人工交互创建三分图的方式无法提取非连通前景区域的问题,提出一种新的交互算法。该算法将CT图像中组织的灰度特征引入颜色估计模型,使用自适应阈值法获取具有相同物理意义但是非连通的组织区域,降低了交互复杂度,提高了交互有效性。实验证明,该法简便高效,解决了传统算法不能应用于非连通组织分割的难题。更多还原

【Abstract】 Digital matting has solved the problem of how to separate the foreground color from the background on the single pixel, and its application is wide. To solve the problem that the tranditional interactive way can not extract a disconnect region, a new algorithm is proposed. Take the grayscale information of CT image into the consideration of the color estimate model, and use an adaptive thresholding to get a disconnect but same properties region. Experimental results show the easiness and efficiency of the method and the application in disconnected tissues segmentation is in effect, which is a impossible task to the traditional approach.更多还原