基于医学图像分割的半自动ROI提取

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。

Semi-automatic ROI Extraction Based on Medical Image Segmentation

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.