Development of a Novel UDL-Model for Dental Abnormality Detection from Digital Photographs

Individuals with Dental Abnormality (DA) have minor to severe health difficulties; therefore, prompt detection and treatment are essential. A typical method for diagnosing DA at the clinical level is image-supported screening. The goal of this project is to create a unique Unified Deep-Learning (UDL) approach for caries identification in teeth from non-invasive digital pictures. This scheme consists of the following stages: (i) taking pictures with a digital camera; (ii) resizing and processing the images; (iii) putting the UDL model for tooth and caries identification into practice; and (iv) verifying and evaluating the scheme's performance. In this work, the Region Proposal Network (RPN) is integrated with the pre-trained MIDNet18 to create the UDL scheme. The effectiveness of this UDL model is tested against the current segmentation strategy, and its effectiveness is validated by the obtained outcomes. The main benefit of UDL-based detection is that the suggested method is easy to use and effective in identifying dental caries in the selected digital photos. Furthermore, when compared to other sophisticated image capture and evaluation methods already in use in dental clinics, the adopted UDL model is straightforward and trustworthy.