With the rapid development of computer science, the application of artificial intelligence (AI) in the field of medical imaging has become increasingly extensive. The temporomandibular joint (TMJ) is structurally complex, with a high incidence of related disorders and diverse clinical manifestations. This review analyzes the current state of research on AI in TMJ imaging diagnosis. Deep learning models based on U-Net and its derivatives have demonstrated outstanding performance in segmentation of condyle and articular disc. Various object detection and feature extraction algorithms have shown excellent diagnostic efficacy for common conditions, such as osteoarthrosis and disc displacement, with some models even achieving expert-level performance on test datasets. Meanwhile, explainable AI provides intuitive justification for model decisions through techniques such as heatmap visualization. Notably, current studies still face critical challenges, including coverage of disease spectra, integration of multimodal data, and model generalizability. Future studies should focus on developing integrated systems that combine diagnosis, segmentation, generation, and interpretation functions. Through multicenter data validation and algorithmic optimization, these efforts will enhance the clinical applicability and decision transparency of models, ultimately laying the foundation for precise imaging diagnosis and intelligent management of TMJ disorders.

In recent years, the application of artificial intelligence (AI) in the medical field, especially in dental imaging diagnosis, has developed rapidly. This review introduces research on AI in detail from the aspects of identifying caries, periapical lesions, vertical root fracture, periodontitis, maxillary supernumerary teeth and impacted mandibular third molars, oral and maxillofacial cystic lesions and temporomandibular joint disorders. Studies have shown that the diagnostic accuracy of AI in the abovementioned oral diseases is equivalent to or even better than that of dentists, indicating that AI has potential in oral disease diagnosis. However, AI models are limited by manual annotation accuracy, poor interpretability, weak generalization ability and difficulty in incremental learning. Future investigations should focus on the development and training of algorithms to improve AI accuracy in oral disease diagnosis.

Cone beam computed tomography (CBCT) has been widely used in various fields of dentistry. The diagnosis of root fractures, especially vertical root fractures (VRFs) with CBCT images, has been a research hotspot since then. Research on this area mainly includes the following five aspects: ①the diagnostic efficiency of CBCT images for root fractures; ②the influence of scanning parameters on the diagnostic accuracy of CBCT images in root fractures, such as scanning field of view, spatial resolution, tube current and tube voltage; ③whether the application of image postprocessing techniques, especially metal artifact reduction (MAR), can improve the diagnostic accuracy of root fractures after root canal treatment and/or there is a post core in the root canal; ④establishment and validation of clinical diagnosis model for vertical root fracture; and ⑤application of artificial intelligence technology and contrast agent in root canals for the diagnosis of CBCT image in root fractures. Compared with periapical radiographs, CBCT images can improve the diagnostic accuracy of root fractures in nonendodontic treated teeth; however, for teeth that have undergone endodontic treatment, the diagnosis of VRF must be combined with clinical signs. Vertical bone resorption in the buccolingual (palatal) direction is a characteristic indicator of VRF. The width of the VRF is an important factor affecting the diagnostic accuracy, but the voxel size used in CBCT scanning is not a necessary factor affecting its diagnostic accuracy; the fracture direction does not affect the diagnostic accuracy of the VRF. Image postprocessing techniques, especially MAR, cannot improve the diagnostic accuracy of VRF and may also reduce the diagnostic efficiency, so they are not recommended for clinical application.