Robust arch detection and tooth segmentation in 3D images of dental plaster models

We present an automated method for determining the dental arch form, detecting the interstices between teeth, and segmenting the posterior teeth in 3D images of dental plaster models. The dental arch form is obtained by a robust two-step curve fitting method that can handle dental models with not only well-aligned but also malaligned teeth. The interstices between teeth are detected by searching for valleys along the dental arch form. We employ a FIR band-pass filter to facilitate the valley detection. Plan-view and front-view range images are utilized for the detection of teeth interstices. The posterior teeth are segmented by tracing edges using the inner product of gradient vectors.