【中图分类号】R445.1 【文献标识码】C【文章编号】1672-5085(2008)03-0164-02

    【摘要】图像分割对于医学图像处理非常重要，只有准确地从图像中提取出信息，才能保证医学系统的可靠性，并为三维重建打下基础。但由于超声图像的形成过程、图像内容的组织性质，以及干扰等因素，使得超声医学图像总是含有很强的噪声。相对与其他类型的图像来说，超声图像的分割要困难得多，本文提出一种新的分割方法，实现分割过程的自动化，给出了肝脏类超声图像的分割实验结果。

    【关键词】超声医学图像；图像分割；峰谷法；Hilbert 曲线；三次样条插值；区域标记

    Shallowly discussed how automatically divides the supersonic medicine picture LiuYan

    【Abstract】 The picture division regarding the medicine picture processing count for much, only has accurately withdraws the information from the picture, can guarantee the medicine system the reliability, and builds the foundation for the three dimensional reconstruction. But as a result of the supersonic picture forming process, the picture content organization nature, as well as factor and so on disturbance, enables the supersonic medicine picture always to include the very strong noise. Is opposite said with other types pictures that, the supersonic picture division must be much more difficult, this article proposed one new division method, the realization division process automation, has produced the liver class supersonic picture division experiment result. ......