A new mathematical modelling based shape extraction technique for Forensic Odontology

Area of Science:

• Forensic Odontology
• Biometrics
• Image Analysis

Background:

• Forensic Odontology is crucial for identifying deceased individuals, especially in mass fatality incidents.
• Current identification methods often rely on comparing postmortem (PM) and antemortem (AM) records, including dental radiographs and photographs.

Purpose of the Study:

• To introduce a novel mathematical algorithm for enhanced person identification in forensic odontology.
• To integrate photographic data with dental radiographs for more robust identification.

Main Methods:

• Utilized the Isoperimetric graph partitioning method for extracting dental image shapes.
• Employed shape matching techniques comparing AM and PM dental images using similarity and distance measures.

Main Results:

• Experimental results indicate that distance metrics provide higher matching accuracy than similarity measures.
• The proposed algorithm demonstrated a high hit rate when utilizing distance-based performance evaluation.

Conclusions:

• The developed algorithm is well-suited for forensic odontologists to accurately identify individuals.
• Distance-based measures are more effective for forensic dental identification using image analysis.