A framework for estimating probability of a match in forensic bite mark identification.

Area of Science:

• Forensic Science
• Dental Forensics
• Biometrics

Background:

• Bite mark analysis in forensic dentistry traditionally relies on subjective human expert comparison.
• Unlike DNA analysis, bite mark comparison lacks a quantitative basis for assigning match probabilities.

Purpose of the Study:

• To propose a framework for empirically estimating the probability of bite mark matches.
• To introduce a quantitative approach to forensic bite mark identification.

Main Methods:

• Collected 3D dental casts and bite mark images from a dental population.
• Utilized image analysis, specifically chamfer distance, to quantify the degree of match.
• Performed logistic regression analysis on population data to estimate match probability based on the degree of match.

Main Results:

• The developed model demonstrated high accuracy in predicting matches and mismatches.
• The model correctly identified 35 out of 42 actual matches.
• The model correctly identified 585 out of 588 non-matches, indicating strong predictive power.

Conclusions:

• The proposed framework offers a quantitative method for estimating bite mark match probabilities.
• This approach enhances the objectivity and reliability of forensic bite mark analysis.
• The methodology has potential applications in other forensic fields requiring quantitative probability assignments.