Bone contour tracing in digital X-ray images based on adaptive thresholding

Abstract

An orthopedic X -ray captures bone images along with surrounding flesh and muscle components. Segmentation of the bone component with a sharp contour is a challenging task as the bone and flesh regions often have pixels with overlapping intensity range. In this paper, we propose a new technique of contour extraction by integrating an entropy-based segmentation approach with adaptive thresholding. The method eliminates the shortcomings of earlier derivative or deformable model based approaches, and can be fully automated. Experiments with several digital X -ray images reveal encouraging results especially for long-bone X -ray images. © Springer-Verlag 2013.