Segmentation of fracture femoral digital radiographs based on grayscale fluctuations and C-V model

Fracture femoral digital radiographs have the characteristics of complicated topology and inhomogeneous intensity distribution. These characteristics make it difficult to segment images effectively. To address this issue, this paper proposes an active contour model combined with grayscale fluctuation information. First, the concept of gradation fluctuation is introduced, and a grayscale fluctuation curve is obtained in the horizontal direction of images. Second, grayscale fluctuation transformation on images is performed according to the monotone interval. Last, the new active contour model is obtained by redefining the level set function, and the inhomogeneous intensity image segmentation is realized accordingly. Experimental results show that the proposed model effectively segments the fracture femoral digital radiographs with high accuracy.