Morphological segmentation of x-ray images

X-ray images are the essential aiding means in clinical diagnosis of fracture all along. According to the characteristic of X-ray images, a novel improving strategy based on homotopy modification of gradients is proposed. Opening-by-reconstruction and closing-by-reconstruction are used to smooth the images. Top-hat transformation and bottom-hat transformation are used to enhance contrast. Combination of top-hat-reconstruction and opening-by-reconstruction is employed to remove complex muscle background. Extended minima transformation, area opening operation, image addition, complement of image, distance transformation and watershed transformation are together employed to compute internal and external markers. Using the markers to modify gradients ensures that the regional minima exactly occur in the bones objects and the muscle background. Correct watershed ridge lines are obtained by applying watershed transform to the modified gradients. Experimental results show that a good segmentation effect can be achieved using the proposed algorithm, and the algorithm is suitable for segmenting X-ray images in interactive fracture diagnosis systems.