This study introduces an automated system employing deep learning-based object detection to identify dental caries from dental radiographs, leveraging advancements in self-supervised learning, particularly masked image modeling (MIM). We emphasize the benefits of self-pre-training, particularly in scenarios where acquiring suitable pre-training data is challenging. Our proposed Masked Deep Embedding of Patches (MDEP) pre-training method, an extension of the masked autoencoder (MAE) transformer, significantly improves dental caries detection performance compared to traditional baselines. Key contributions include utilizing self-supervised learning for limited dental radiograph data, replacing MAE’s patch reconstruction with the reconstruction of patch embeddings, and achieving notable performance enhancements in caries detection. This research paves the way for AI-driven solutions in dentistry, offering valuable tools for dentists and patients in the early detection and management of dental caries. Additionally, we validate the robustness and generalizability of our proposed method using another dataset, CariesXrays. Our findings indicate that MDEP achieves higher precision and recall rates, providing a significant improvement in diagnostic accuracy. The implications of this study extend beyond dental caries detection, suggesting that self-supervised learning can be effectively applied to other areas of medical imaging where data scarcity is a significant challenge.

This study proposes an automated system using deep learning-based object detection to identify implant systems, leveraging recent progress in self-supervised learning, specifically masked image modeling (MIM). We advocate for self-pre-training, emphasizing that its advantages when acquiring suitable pre-training data is challenging. The proposed Masked Deep Embedding (MDE) pre-training method, extending the masked autoencoder (MAE) transformer, significantly enhances dental implant detection performance compared to baselines. Specifically, the proposed method achieves a best detection performance of AP = 96.1, outperforming supervised ViT and MAE baselines by up to +2.9 AP. In addition, we address the absence of a comprehensive dataset for implant design, enhancing an existing dataset under dental expert supervision. This augmentation includes annotations for implant design, such as coronal, middle, and apical parts, resulting in a unique Implant Design Dataset (IDD). The contributions encompass employing self-supervised learning for limited dental radiograph data, replacing MAE’s patch reconstruction with patch embeddings, achieving substantial performance improvement in implant detection, and expanding possibilities through the labeling of implant design. This study paves the way for AI-driven solutions in implant dentistry, providing valuable tools for dentists and patients facing implant-related challenges.