To deal with the extreme over-segmentation problem of region-based watershed methods,two approaches have been proposed. Firstly,we perform the watershed transform on the floating-point moving image in stead of the morphological grad image,which can locate the contour more accurately.
针对基于区域的分水岭分割算法中通常存在的严重过分割现象,通过两点措施来改进:其一,以梯度图像的浮点活动图像取代梯度图像进行分水岭变换,使边缘定位更准确;其二,以基于面积和对比度控制的合并小区域准则,有效抑制过分割现象,同时满足感兴趣区的分割要求。
参考来源 - 基于区域生长法的医学图像分割研究·2,447,543篇论文数据，部分数据来源于NoteExpress
An adaptive marker extraction-based watershed algorithm is proposed to overcome the over-segmentation problem.
针对分水岭算法过分割问题，提出一种基于自适应提取标记的改进算法。
To solve the problem of the over-segmentation, the criterions of region intensity and edge strength and region size are combined.
为解决过度分割的问题，给出了区域的灰度特性和边界强度及区域尺寸相结合的融合准则。
But the traditional watershed transformation has a serious problem of over segmentation, and noise of the image and fake textures may submerge the wanted edge information.
但传统分水岭变换过分割问题严重，图像的噪声和虚假纹理会淹没真正想得到的边缘信息。
应用推荐