Related Experiment Video
Updated: Apr 25, 2026

09:49
Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
Published on: December 24, 2015
16.2K
Assessment of presentation methods for ReFace computerized facial approximations
Adam H Richard1, Connie L Parks1, Keith L Monson2
1Counterterrorism and Forensic Science Research Unit, Visiting Scientist Program, Federal Bureau of Investigation Laboratory Division, Quantico, VA, USA.
Forensic Science International
|August 17, 2014
Summary
Presenting multiple facial approximations improves recognition accuracy. Participants were better at identifying a correct approximation among several options, suggesting their value as investigative tools.
Area of Science:
- Forensic Science
- Psychology
- Computer Graphics
Background:
- Facial approximations are crucial investigative tools for identifying missing persons.
- Current presentation methods lack standardized effectiveness metrics.
- Previous accuracy assessments may be unsuitable for approximation's investigative role.
Purpose of the Study:
- To determine the most effective presentation method for facial approximations.
- To evaluate a novel method for assessing facial approximation accuracy.
Main Methods:
- Five distinct presentation methods for facial approximations were tested.
- Participant recognition accuracy was measured for each method.
- A new assessment technique for approximation accuracy was evaluated.
Main Results:
- Presentation methods using multiple images of an approximation generally improved performance.
- Optimal presentation method varied across different approximation subjects.
- Participants were more successful identifying correct approximations from a selection than a single choice.
Conclusions:
- Facial approximations may be more effective as investigative tools than previously recognized.
- Multiple-image presentation formats show promise for enhancing recognition.
- Further research is needed to refine assessment methods and presentation strategies.

