Automatic segmentation of cervical vertebrae in X-ray images

Physiological parameters of vertebrae are important for cervical condition assessment. In order to measure the parameters fast and accurately, automatic segmentation instead of manual key point placement has become an imperative for diagnosing. We propose an applicable automatic segmentation system for medical image of cervical spine. The system includes a series of algorithms: a parallel cascade structure based Haar-like features and the AdaBoost learning algorithm used to detect the location of cervical vertebrae as a initial position of Active Appearance Model (AAM), multi-resolution AAM search applied to improve the speed and accuracy of AAM fit, and combination of global AAM and local AAM used to achieve more effective matching of details of vertebrae. Experiments on the cervical spine databases show a significant increase in speed, robustness and quality of fit compared to previous methods.