Next-generation osteometric sorting: Using 3D shape, elliptical Fourier analysis, and Hausdorff distance to optimize...

Area of Science:

• Forensic Anthropology
• Bioinformatics
• Computer Vision

Background:

• Pair-matching bones is crucial for sorting commingled human remains.
• Traditional methods rely on visual inspection and manual measurements.
• Computational analysis offers a more objective and efficient approach.

Purpose of the Study:

• To develop and evaluate a fast 3D computerized shape-analysis method for whole-bone pair-matching.
• To assess the method's accuracy and efficiency in identifying bilateral bone pairs.

Main Methods:

• Utilized a 3D elliptical Fourier analysis function to create a shape signature from bone perimeters.
• Reduced 3D point cloud data to 100 points while preserving shape information.
• Applied mean Hausdorff distance to compare mirrored bone outlines for pair-matching.

Main Results:

• Achieved high true positive rates: 1.00 for femora, 1.00 for humeri, and 0.92 for tibiae.
• The lowest mean Hausdorff distance proved to be the most effective pair-match criterion.
• Computational time for 469 pairwise comparisons was only 5 seconds.

Conclusions:

• The proposed 3D computerized shape-analysis method is fast, accurate, and viable for real-world forensic applications.
• This technique significantly enhances the efficiency of sorting commingled skeletal remains.