Purpose: The purpose of this study was to investigate the efficiency of panoramic radiography, conventional (cross-sectional) tomography, and computerized tomography for location of the mandibular canal before implant placement in the posterior region of the mandible. Materials and Methods: Edentulous mandibles from 6 dry adult human skulls were used in this study. Four measurements (D1, D2, D3, D4) were made of 12 areas, one on each side of each mandible. Panoramic radiographs, conventional tomograms, and computerized tomograms were obtained. On each image, measurements were made for localization of the mandibular canal by one researcher. All measurements were repeated 3 times within a period of 3 weeks. Upon completion of imaging, the mandibles were surgically sectioned to provide direct measurements. The measurements obtained from the images were compared with direct measurements. Pearson correlation coefficients were calculated to detect statistical correlations between repeated measurements. The Dunnett t test was performed for statistical comparison of measurements from images and direct measurements. Results: Pearson correlation coefficients showed strong linear correlation for all measurements (P .01). No statistically significant difference was observed between direct measurement and D1, D2, or D4 (P .05), but a statistically significant difference for D3 (buccolingual width 5 mm under mandibular crest; Dunnett t test; P > .05) between measurements was obtained from the images and direct measurements. Conclusion: The measurements obtained from computerized tomographic images were more consistent with direct measurements than the measurements obtained from panoramic radiographic images or conventional tomographic images. Keywords: dental implant, imaging methods, mandibular canal, treatment planning