Hierarchical Interactive Image Segmentation using Irregular Pyramids

In this paper we describe modifications of irregular imagesegmentation pyramids based on user-interaction. We first build a hierarchyof segmentations by the minimum spanning tree based method,then regions from different (granularity) levels are combined to a final(better) segmentation with user-specified operations guiding the segmentationprocess. Based on these operations the users can produce a finalimage segmentation that best suits their applications. This work can beused for applications where we need accuracy in image segmentation, inannotating images or creating ground truth among others.