Mandibular nerve canal identification for preoperative planning in oral implantology

The identification of relevant anatomies is fundamental for correct preoperative planning in oral implantology. In this setting, few state-of-the-art software products have the option of drawing the mandibular nerve canal, and almost none has a related automatic recognition functionality. We present an algorithm for the automatic recognition and drawing of the mandibular canal for oral implantology 3D-based software systems. The developed algorithm uses two user-identified extremities of the canal to determine a checkpoint at the mandibular foramen ascent base. It identifies the mandibular canal region moving within the radiographic volume, exploiting adaptive thresholds and ROI-intersection-based movement vectors. Finally, the software uses a Catmull-Rom-spline-based drawing feature to produce the canal 3D mesh. Experimental evaluation on scans of 7 human mandibles resulted in 13 successful identifications out of 14, while 1 failed due to extreme scattering.