PROJECT TITLE :

Tufts Dental Database A Multimodal Panoramic X-Ray Dataset for Benchmarking Diagnostic Systems

ABSTRACT:

Because there is such a wealth of clinical data, both image-based and non-imagery-based, the implementation of artificial intelligence in dental care holds a lot of potential for the future. Dental radiographs, when analyzed by a trained professional, can provide vital information that can aid in clinical diagnosis and treatment. In recent years, Convolutional Neural Networks have achieved the highest accuracy possible in a variety of benchmarks. One of these benchmarks is the analysis of dental X-ray images to improve the quality of clinical care. In this paper, we present the Tufts Dental Database, a brand-new X-ray panoramic radiography image dataset. The database was created by Tufts University. This dataset includes one thousand panoramic dental radiography images that have been labeled by a specialist to indicate any abnormalities or missing teeth. The radiography images were classified based on five distinct levels, which were the anatomical location, peripheral characteristics, radiodensity, effects on the surrounding structure, and abnormality category, respectively. This multimodal dataset is the first of its kind, and it also includes the radiologist's expertise, which was recorded using an eye-tracking and think-aloud protocol. A publicly available dataset that can assist researchers in incorporating human expertise into AI and achieving more robust and accurate abnormality detection is one of the contributions of this work. Other contributions include 1) a benchmark performance analysis for various state-of-the-art systems for dental radiograph image enhancement and image segmentation using deep learning; and 2) an in-depth review of various panoramic dental image datasets, along with segmentation and detection systems. The release of this dataset has the goals of accelerating the development of AI-powered automated abnormality detection and classification in dental panoramic radiographs, improving tooth segmentation algorithms, and enabling radiologist expertise to be distilled into AI.