Long-bone fracture detection in digital X-ray images based on concavity index

Abstract

Fracture detection is a crucial part in orthopedic X-ray image analysis. Automated fracture detection for the patients of remote areas is helpful to the paramedics for early diagnosis and to start an immediate medical care. In this paper, we propose a new technique of automated fracture detection for long-bone X-ray images based on digital geometry. The method can trace the bone contour in an X-ray image and can identify the fracture locations by utilizing a novel concept of concavity index of the contour. It further uses a new concept of relaxed digital straight line (RDSS) for restoring the false contour discontinuities that may arise due to segmentation or contouring error. The proposed method eliminates the shortcomings of earlier fracture detection approaches that are based on texture analysis or use training sets. Experiments with several digital X-ray images reveal encouraging results. © 2014 Springer International Publishing.