A multi-factor knuckle and nail bed verification tool for forensic imagery analysis.

Area of Science:

• Forensic Science
• Computer Science
• Biometrics

Background:

• Child sexual exploitation often involves grooming through explicit images or videos.
• Offenders may use hand, knuckle, and nail imagery to conceal identity.

Purpose of the Study:

• To develop a novel biometric hand verification tool for identifying online child sexual exploitation offenders.
• To authenticate hand imagery against a reference database using advanced image processing and machine learning.

Main Methods:

• Experiments were conducted on two hand datasets (Purdue University and Hong Kong).
• System performance was evaluated based on parameters like camera distance, orientation, and model architecture.

Main Results:

• Optimal performance was achieved when images were sampled from the same database under identical capture conditions.
• The study explored various parameters influencing biometric verification model reliability.

Conclusions:

• The biometric hand verification tool provides a robust solution for law enforcement.
• This technology can enhance the effectiveness of investigations into online child sexual exploitation offenders.