ÀÏÍõÂÛ̳

• Mortimer Cavanah

Mortimer Cavanah

Advisor: Maja Å ±ðÅ¡±ð±ôÂá

Exploring the Applications of Deep Learning in Dental Anthropology

The identification of dental growth stages in immature individuals is widely used to estimate chronological age and/or assess skeletal maturity in many academic and professional fields (e.g., forensic anthropology, bioarchaeology, biomedicine). The current process used for identifying these dental stages involves individual assessors hand-scoring dental development based on non-metric, qualitative criteria-- consuming large amounts of both time and human labor. Much of this time and labor consumption could be reduced by using deep learning techniques to develop a program capable of identifying dental growth stages. By using Google’s deep learning framework TensorFlow, this research project will focus on developing and curating a dataset of jaw radiographs and then utilizing this dataset to train a deep learning algorithm to recognize dentition and dental growth stages via supervised learning. This project is an exploration into determining how to make qualitative methods of scoring ‘digestible’ and recognizable to artificial intelligence in order to produce results consistent with human scoring, while ideally reducing time and labor costs, as well as improving reproducibility and accuracy.