A study on artificial intelligence-based identification of bone fenestration and dehiscence in CBCT images

The identification of orthodontic bone fenestration and dehiscence holds great importance in the field of oral orthodontics. They can affect the outcomes of orthodontic treatment and the oral health of patients. This study presents the development of an artificial intelligence-based method for the automatic identification of orthodontic bone fenestration and dehiscence using high-resolution Cone Beam Computed Tomography (CBCT) images. We propose a novel methodological framework to accurately recognize these features in high-resolution CBCT images. We have validated the effectiveness of this method using clinical oral CBCT images. The results demonstrate that our method achieved significant improvement in the identification of bone fenestration and dehiscence. The accuracy rate is 89.50%, recall rate is 92.42%, and F1-score is 93.13%. Compared to traditional methods, our approach shows great advantages in identification accuracy and efficiency. It strongly supports the early diagnosis and treatment of orthodontic bone fenestration and dehiscence. This study not only provides a new intelligent diagnostic tool for the field of oral orthodontics but also offers new ideas and methods for future oral imaging research based on artificial intelligence. With further model optimization and expansion of clinical datasets, we expect to enhance the accuracy and robustness of this method and promote its widespread application in clinical practice.