Journal of Prosthodontics News
Feasibility and Accuracy of Single Maxillary Molar Designed by an Implicit Neural Network (INN)-Based Model
Now online in the Journal of Prosthodontics, a study evaluating the use of a novel implicit neural network (INN)-based model for maxillary molar reconstruction co-authored by ACP member, Wei-Shao Lin, DDS, PhD, MBA.
Creating a comprehensive physical or virtual prosthetic wax-up is essential for optimal implant planning. However, despite the capabilities of modern CAD software, substantial manual inputs and personalized adjustments remain necessary to achieve the desired position, occlusion, and esthetics.
Artificial intelligence (AI) has shown the potential to improve this process. Deep learning (DL) is an advanced AI-based machine learning technique rooted in neural networks. It can be trained to process data effectively and is being applied in implant prosthodontics. To evaluate the feasibility of using a neural network (INN)-based mode, a total of 500 sets of full dental scans containing intact right and left maxillary first molars (#3 and #14) and adjacent teeth were included in this study. The digital maxillary casts were duplicated: one set served as the original, while the other had one maxillary first molar removed. Two INN-based models were developed, including the point convolution for surface reconstruction (POCO-only) model and the POCO with point multilayer perceptron (POCO-PointMLP) model. Each model was trained with either 12,000 or 50,000 sampling points from the training dataset.
The novel INN model demonstrated strong predictive capabilities as evidenced by key evaluation metrics such as CD, F-score, and IoU values after training. The model achieved its best performance metrics when utilizing a combination of the POCO and PointMLP modules with a sampling point count of 50,000. This proposed INN model successfully generated a single crown surface for maxillary molars, yielding comparable morphology discrepancies to those designed by skilled technicians.
Wang Y, Shi Y, Li N, Lin W-S, Tan J, Chen L. Feasibility and accuracy of single maxillary molar designed by an implicit neural network (INN)-based model: A comparative study. J Prosthodont. 2025; 1–8. https://doi.org/10.1111/jopr.14035
Previous
Next
Back to News