Neck pain has great economical consequences but due to the relationship between intervertebral motion and the functionality of the spine, investigating the rotation and translation of adjacent vertebrae can be used to determine the underlying cause. Fluoroscopic imaging can, with the high number of images taken, provide a more accurate reflection of the intervertebral motion than other imaging modalities. The estimation of the rotation and translation most often depends on the localization of four landmarks, which are usually the corner points of the vertebral bodies in the cervical spine. The detection of the landmarks is usually based on manual markings, which can be a tedious process. The aim of the present study was to detect the landmarks of the cervical vertebrae C2-C7 in fluoroscopic recordings. Four subjects performed extension and flexion of the neck while wearing glasses with external markers for the occiput. The center points of the external markers and the four landmarks of C3-C6 along with the inferior corner points of C2 and superior corner points of C7 were automatically detected. The landmarks were detected using a multisegmentation approach and the automatic detection method was tested on five frames of each fluoroscopic recording. The external markers were detected correctly in 27 of the 40 tested frames while the corners of the cervical vertebrae were detected in 31 frames. The method shows, that it is possible to detect the cervical vertebrae without manual interaction.