中文摘要：

介绍了Mumford-Shah模型的基本理论,将众多的Mumford-Shah模型改进算法归纳为内嵌改进型和顺接改进型两种,并指出内嵌改进型主要改进了Mumford-Shah模型的计算方式,顺接改进型是以串行的方式融合了Mumford-Shah模型和其它方法。最后指出Mumford-Shah模型仍需要改进,以解决复杂或多目标图像的分割问题。

英文摘要：

Mumford-Shah models are the popular research in the fields of image processing and computer vision.Firstly the theories are presented.Then all the improving algorithms are divided into two categories,i.e.embedded models and sequent ones.Embedded models mainly improve the inner processing of Mumford-Shah models and sequent ones integrate Mumford-Shah models with other al-gorithms in the serial mode.Finally,their applications are mentioned.More amelioration for Mumford-Shah models to resolve the problems on complex image and multi-object image.