Related Experiment Video
Updated: Aug 30, 2025
10:01
The Visual Colorimetric Detection of Multi-nucleotide Polymorphisms on a Pneumatic Droplet Manipulation Platform
Published on: September 27, 2016
7.7K
A PNU-Based Methodology to Improve the Reliability of Biometric Systems
Paola Capasso1, Lucia Cimmino1, Andrea F Abate1
1Department of Computer Science, University of Salerno, 84084 Fisciano, Salerno, Italy.
Sensors (Basel, Switzerland)
|August 26, 2022
Summary
This study introduces a hybrid face recognition method robust to face masks and deep fakes. The system uses Source Camera Identification (SCI) with Pixel Non-Uniformity (PNU) to enhance biometric security.
Area of Science:
- Computer Science
- Biometrics
- Image Analysis
Background:
- Face recognition systems face challenges due to COVID-19, including mask-wearing and contactless acquisition requirements.
- Existing biometric systems show reduced reliability with occluded faces and are vulnerable to deep fake attacks.
Purpose of the Study:
- To develop a hybrid methodology enhancing face recognition reliability.
- To address challenges posed by face masks and deep fake alterations in biometric security.
Main Methods:
- A hybrid methodology integrating Source Camera Identification (SCI) based on Pixel Non-Uniformity (PNU).
- Analysis of video stream integrity and detection of tampered/fake frames.
- Prototype implementation for real-world use-case evaluation.
Main Results:
- Achieved high accuracy in face recognition despite subjects wearing face masks.
- Successfully detected deep fake video alterations, generating multiple alert messages.
- Demonstrated robustness against mask-induced occlusions and enhanced reliability against forgery attacks.
Conclusions:
- The proposed hybrid methodology significantly improves face recognition system reliability.
- The system is effective in overcoming mask-related occlusions and preventing deep fake manipulations.
- This approach offers a more secure and robust solution for modern biometric security needs.
