This study compared the accuracy of new, FDA-approved, image-analysis software to conventional radiographic assessment techniques for the measurement of intervertebral motion. Six adult human cadaveric lumbar spines (L1-S1) were individually mounted in a custom Plexiglas device and electromagnetic sensors were rigidly mounted to the spinous processes of L3, L4, and L5. Lateral radiographs of the spines in neutral, full flexion, and full extension were digitized and analyzed both using the software and manually by three orthopedic surgeons. Compared to intervertebral rotations determined from the electromagnetic device, the errors in rotations reported by the software and surgeons were 0.47+/-0.24 degrees and 2.16+/-0.78 degrees , respectively. Rotations measured by the surgeons were significantly less accurate and more variable than that of the software (p<0.05).