改进分水岭算法在医学图像分割中的应用

医学图像存在病变区域和背景区域,病变区域是分割的重点,亦印感兴趣区域.针对传统的分水岭变换由于对噪声和细密纹理的敏感会产生严重的过分割现象,提出两点改进:首先对原始图像运用高频强调滤波作图像增强预处理,补偿图像轮廓、突出图像的边缘信息以使图像显得更加清晰;其次是采用修改了标记符的梯度图像进行分水岭分割,整个分割过程无须进行再进行后续的合并处理,降低了分割的复杂性.仿真实验结果表明,这种改进的方法不但使过分割现象得到了很好的抑制,而且医学图像中的病变区域被有效分割出来,同时该方法还支持任意形状,且与JPEG2000标准兼容,分割效果很好.

Application of Medical Image Segmentation Based on Improved Watershed Segmentation Algorithm

There are pathological change region and background in the medical images.The pathological change region is the emphasis of image segmentation.In this paper,two methods have been introduced to resolve the over-segmented of watershed algorithm because of the sensitivity of noise and fine texture in the region-based segmentation process: firstly,the high-frequency emphasis filter is used in the original image for enhancement preprocessing,which can make up image contour,and also can highlight the image edge i...