Hierarchy of Image Segmentation
Shen Lansun, Wang Aimin (Signal and Information Processing Laboratory, Beijing Polytechnic University, Beijing, 100022)
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. In this article, a new generalized definition for segmentation is presented. A 4-level scheme for generalized segmentation is constructed according to the hierarchy of segmentation purposes. This article also surveys applications and current methods for the 4-level segmentation tasks. The results are beneficial to further research on the applicability of current segmentation algorithms.