Two-Stage Segmentation Method for Context-Sensitive Image Analysis

Abstract

This article considers two-stage segmentation method for context-sensitive image analysis. We represent a combination of two methods: pyramidal segmentation and Grabcut. The results of the pyramidal segmentation are used for the Grabcut as input data. The resulting regions should be used for further detailed recognition. In conclusion shows the comparative results of pyramidal segmentation, Grabcut segmentation and two-stage segmentation. © Springer International Publishing Switzerland 2014.