图像分割研究综述

Study Surveys on Image Segmentation

图像分割是指将一幅图像分解为若干互不交迭区域的集合 ,是图像处理与机器视觉的基本问题之一。本文对常见的分割方法按照数据驱动与模型驱动两大类型进行了综述 ,并介绍了图像分割中的物理模型与随机场模型、半自动分割策略、图像分割的定量评价方法等。

Image segmentation denotes a process by which a raw image is partitioned into nonoverlapping regions.The image segmentation problem is treated as a key issue in image processing and machine vision. The current methods based on two categories, data and model driven segmentation are surveyed.In addition, the other points related to segmentation,including physics based image model,random field image model,semiautomatic segmentation, and quantitative evaluation of segmentation,etc. are reviewed.