A novel approach to identify dental implants or fixtures from orthopantomographs using artificial intelligence custom model

Abstract

This study is done to identify brands and models of dental implants from radiographs and to develop a space invariant artificial neural network (SIANN). The efficiency of artificial intelligence is to identify fixture or dental implant systems from configurations-based orthopantomographs were evaluated. The performance of the model is analyzed by calculating the accuracy and recovery operating characteristics (ROC). AI model recognize the implant type and its model. Success were predicted.AI model for the design identification of dental implant or fixtures. Their prediction has shown enormous potential but are still undeveloped that much. Total 56 Cowell Medi orthopantomographs that consists of three types of dental implant system with similar internal conical connection and shape, were unsymmetrically divided into training and authenticated datasets (76%) and the datasets for evaluation (20%). Based on fine tuning and pretrained space invariant artificial neural network architecture (yolov8) techniques like image preprocessing and transfer learning were executed. Afterwards, evaluated datasets or test datasets were used to calculated the sensitivity, accuracy, specificity. Area under the receiver operating characteristics curve (AUC), receiver operating characteristics curve were compared between space invariant artificial neural networks (SIANN). Hereby, the artificial intelligence model developed that is efficient enough to correctly recognize the type of dental implants or fixtures. The main aid of it is to established a greater number of datasets within lesser time and to do their cross sectional or transverse analysis. The categorizer evaluated thereby will be able classify dental implants. The average accuracy of this model is 0.84 or 84%.

Ähnliche Arbeiten