An Enhanced YOLOv8 Approach for Quick Human Identification via Bite Marks
Main Article Content
Abstract
Identification of individuals by their dental features or bite mark impressions constitutes the discipline of forensic dentistry. This article describes the development of an automatic system for comparing bite mark images and dental models of suspects. It determines the best match by comparing bite mark textures to a database of dental casts. These results are compared with expert opinions of forensic odontologists. This preliminary study assesses whether YOLOv8(You Only Look Once) is a theoretically feasible method for identifying and interpreting patterns of bite mark texture evidence imaged on pink wax impressions. The agreement with the experimental data is fairly good. The effectiveness of YOLOv8 for detection and recognition of complex textural patterns in bite marks has great potential to enhance the rapidity and reliability of human identification, providing a hopeful path for the development of forensic odontology.