In this paper we propose a new method of vertebrae segmentation in medical images. X-ray images of the spinal columns are analysed in order to extract a closed contour for each vertebra. To achieve this goal, we proceed by steps: the starting one is a segmentation approach based on the selection of each vertebra region. We use these regions information to identify each individual vertebra by its contour. For this task, we propose a polar signature representation of the contour using the image gradient of each region. Lastly, an edge closing method exploiting polynomial fitting is used.

We use the resulting extracted contours to determine vertebral mobility. For this, we proceed by the extraction of some parameters characterizing each vertebra, related to their form, position and orientation.