Options
2014
Conference Paper
Titel
Automatic classification of salient boundaries in object-based image segmentation
Abstract
We present a supervised classification approach for image segmentation that operates in an object-based image representation and combines object features with boundary features. While classical algorithms focus on either regions (i.e. objects) or edges (i.e. boundaries), we offer a hybrid solution that takes both aspects into consideration. To illustrate the capacity of this approach, we apply the proposed classification to CT bone segmentation.