Landmark-less registration of craniomaxillofacial images acquired for 3D photography and CBCT

Superimposition of craniomaxillofacial images acquired by cone-beam computed tomography and three-dimensional photography can help assist diagnosis and surgical planning. Conventional approaches identified prominent facial landmarks on both modalities and assessed their correspondence. Considering, however, variation of facial expressions when imaged at different timing, landmark registration is challenging and can be inappropriate. This paper proposes a disturbance-region removal procedure to improve the efficacy of registration. The disturbance regions are defined as those exhibiting strong responses in the concavity intensity maps that are computed from the facial surface mesh. An adapted symmetric region growing algorithm follows to form connected disturbance regions that are removed prior to superimposition of both modalities. The results show a 28% better match of overall correspondence of the facial fiducial markers. Instead of being the registration guides in conventional approaches, in this study the fiducial markers are employed as a means to assess the performance of registration.