Conventional and Novel Methods for Facial-Image Identification.

Area of Science:

• Forensic Science
• Biometrics
• Image Analysis

Background:

• Facial image identification is crucial in forensic investigations.
• Conventional methods include morphological comparison, anthropometric indices, and face-to-face superimposition.
• These traditional methods have varying degrees of reliability and application in casework.

Purpose of the Study:

• To review conventional and novel facial identification techniques.
• To introduce and evaluate a new 2D/3D facial image superimposition method.
• To propose a 3D facial image database retrieval system for suspect screening.

Main Methods:

• Review of established forensic facial identification techniques.
• Introduction of a novel 2D/3D facial image superimposition method.
• Discussion of reliability assessments and expert opinions in facial identification.

Main Results:

• Conventional methods are ranked by casework frequency: morphological comparison, anthropometric indices, and face-to-face superimposition.
• A novel 2D/3D superimposition method is presented with its pros and cons.
• Reliability of facial identification, including expert judgment, is assessed.

Conclusions:

• Facial identification methods range from traditional comparisons to advanced 2D/3D superimposition.
• A new 3D facial image database retrieval system offers potential for efficient suspect screening.
• Enhancing the reliability and efficiency of facial identification remains a key forensic challenge.