Tooth Enamel segmentation from dental x-ray using MASK R-CNN Algorithm

In order to detect cavities on teeth, dentists analyse dental x-rays. Initially, dental cavities develop on the delicate surface of the teeth, known as enamel, and subsequently progress to the pulp and the root of the teeth. The reconstruction of three-dimensional dentition models for use in a variety of clinical applications is contingent upon the precise segmentation of teeth. Nevertheless, performing manual segmentation is time-consuming and susceptible to intra- and inter-observer variability. Convolutional neural network (CNN) algorithms can automatically segment teeth; however, the majority of CNN algorithms necessitate manual marking on a vast training database and extreme computational capacity. Theoretical Orientation and Methodology: This investigation assesses the feasibility of employing a transfer-learning algorithm, “Mask R-CNN,” which necessitates a modest training database and minimal computational resources. This work is designed to accomplish our objective of segmenting the enamel region of the teeth from dental x -rays using an algorithm such as Mask renn. The extraction of the enamel region is essential in the process of cavity detection using deep learning, as the cavity initially develops on the enamel region. The “Mask R-CNN” method is implemented with a standard computational capacity and a minimal training database to accomplish high-quality dental recognition within a reasonable train duration. This has the potential to promote more precise treatment planning and diagnosis by eliminating subjective errors.