A multimodal orthodontic dataset

In recent years, orthodontic patients have placed growing attentions on facial changes. Orthodontic examinations have evolved to three-dimensional (3D) multimodality, laying the groundwork for the development of virtual orthodontic patients. Meanwhile, artificial intelligence (AI) assisted techniques can automatically register multimodal data and improve precision. However, there remains a lack of orthodontic multimodal datasets. This dataset presents a multimodal orthodontic dataset comprising 523 data, including reconstructed Cone-Beam Computed Tomography (CBCT) data, partially registered oral scans and face scans. Among these, reconstructed CBCT data without registered multimodal data include dental anatomical landmarks. The data includes adolescents and adults, covering both permanent and mixed dentitions, along with Angle Class I, II, and III malocclusions, and facial types ranging from convex and concave to straight profiles. Providing a technical basis for AI-assisted construction of automated, high-precision, and realistic orthodontic virtual patients.