A Regional Energy Function Based Approach for Fast Tooth Segmentation from CBCT Images

Segmenting 3 D teeth models from cone beam computed tomography images is significantly important for orthodontics. In this paper, we propose a new approach based on a regional energy function to solve the fuzziness problem of adjacent teeth crowns and root-jaw issues. The initial curve is first calculated based on the endodontic seeds and gradients. Then, the moving range of the initial curve is restricted by the intensity distribution. Within this moving range, we propose a regional energy function based on intensity, gradient and smoothness. By optimizing the function, the final curve of the slice can be generated, and is applied as the initial curve of next slice. The final tooth model can be reconstructed by Marching Cubes algorithm. The experimental results on CBCT datasets demonstrate that our method is accurate and efficient for the fuzziness problem in orthodontic treatment.