基于医学图像分割的半自动ROI提取  

Semi-automatic ROI Extraction Based on Medical Image Segmentation

作　　者：周娇[1] 徐向民[1] 吕华琳[1] 

机构地区：[1]华南理工大学电信学院,广东广州510640

出　　处：《电视技术》2007年第2期78-80,共3页Video Engineering

基　　金：广东省自然科学基金资助(31449;05100514);广东省科技攻关项目(2005B10201052)

摘　　要：针对当前ROI提取方法,提出了一种基于图像分割的半自动ROI提取算法,一方面保证提取出的ROI是有意义的对象,另一方面人为指定因素更小,对ROI提取的智能性有较大提高。在图像分割中提出一种基于改进的分水岭算法的医学图像分割方法,引入了浮点活动图像代替梯度图像进行分水岭变换,使分割结果边缘定位更准确,在分水岭变换之后,提出基于面积和对比度控制的合并小区域准则,可得到更好的分割效果。In allusion to current ROI extraction methods, a semi-automatic ROI extraction method based on image segmentation is proposed. On one hand, the extracted ROI is ensured to be meaningful. On the other hand, man-made factors are even smaller, boosting greatly the intelligence in ROI extraction. A medical image segmentation algorithm is proposed based on the improved watershed algorithm. The gradient image is introduced to replace the floating video images to process watershed transformation in the algorithm, making the segmentation margin results more accurate. After watershed transformations, rules for merging small regions are proposed based on the size and contrast control to obtain even better segmentation results.

关 键 词：图像分割 分水岭算法 梯度图像 合并小区域准则 

分 类 号：TN911.73[电子电信—通信与信息系统]