A Novel Algorithm for Forensic Identification Using Geometric Cranial Patterns in Digital Lateral Cephalometric...

Area of Science:

• Forensic Anthropology
• Radiology
• Computer Science

Background:

• Lateral cephalometric radiographs are vital in dentistry and orthodontics.
• Their application in forensic identification, particularly for burned individuals or mass disasters, presents significant challenges.
• Comparing antemortem (AM) and postmortem (PM) radiographs is a key method for identification.

Purpose of the Study:

• To introduce and evaluate a novel algorithm for extracting cranial patterns from digital lateral cephalometric radiographs for identification.
• To assess the algorithm's efficacy in matching PM cephalograms with AM records for accurate individual identification.

Main Methods:

• A novel algorithm was developed to encode cranial patterns from digital lateral cephalometric radiographs into a database.
• The algorithm was validated using pre- and post-treatment cephalograms from an orthodontic archive, simulating AM and PM data.
• An Automatic Error Reduction (AER) function was applied to enhance similarity scores.

Main Results:

• The algorithm achieved high accuracy (97.5%), sensitivity (97.7%), and specificity (95.2%), correctly identifying 350 out of 358 cases.
• The mean similarity score improved significantly from 91.02% to 98.10% after applying the AER function.
• Intra-observer error analysis indicated a low average Euclidean distance of 3.07 pixels for landmark selections.

Conclusions:

• The proposed algorithm demonstrates significant potential for identity recognition using cranial patterns from cephalometric radiographs.
• The method offers a promising approach for forensic identification, particularly in situations where traditional methods are difficult.
• Future enhancements could involve integrating artificial intelligence (AI) algorithms to further improve performance.