Initial investigation of mathematical morphology for the digital extraction of the skeletal characteristics of trabecular bone.

Abstract

To develop a filter utilizing mathematical theory to extract the skeletal patterns of trabecular bone. Studies of morphology in the extraction of patterns of calcification in mammograms provided the theoretical framework. Using these studies as a basis, a morphological filter was applied to extract skeletal patterns from digital images of trabecular bone. Sequential images (subset) were combined in a structured fashion to create an aggregate (sumset) which compared with the original images, skeleton and line skeleton images. Binary images of the skeletal patterns in continuous, round and mesh-like forms were obtained from the original images by processing with the skeleton operation using a disc-shaped single structuring element. The line skeleton operation using line structuring elements with constant directions allowed the extraction of linear and discontinuous patterns. Both the skeleton and line skeleton operations extracted binary subset images which depicted skeletal patterns correlating with the operation sequence. Modification of the morphological filter enhanced the extraction of skeletal characteristics of trabecular bone. A morphological filter may be a useful adjunct in computer-aided structural analysis of bone.