Deidentification of facial images using composites

Area of Science:

• Medical photography
• Facial image analysis
• Patient privacy

Background:

• Maxillofacial surgeons use photography for education and documentation.
• Facial images are highly identifiable, posing privacy challenges.
• Traditional deidentification methods compromise image quality or identity concealment.

Purpose of the Study:

• To develop and evaluate a novel method for facial image deidentification using blended facial composites.
• To assess the efficacy of this method in concealing identity while retaining visual educational components.

Main Methods:

• Twenty composite facial images were created.
• Student subjects viewed the composites, initially unaware they were altered.
• Subjects were later informed about the composite nature and re-evaluated the images.

Main Results:

• No familiar faces were recognized when subjects were unaware of composites (0%).
• Recognition increased significantly (62%) when subjects were primed to the presence of composites.
• Composites were rated as clinically realistic and superior to traditional deidentification methods.

Conclusions:

• Blended facial composites show promise for effective facial image deidentification.
• This technique balances patient privacy with the educational value of facial images.
• Further large-scale studies are recommended to validate these findings.