In this chapter we examined two passive approaches to secure an iris recognition system against insertion attacks by verifying the authenticity of the iris images. The first one, named PSI, is based on the photo response non-uniformity (PRNU) of image sensors and the second one, named ITC, exploits the texture information of unrolled iris images. The examination was performed using images from nine distinct iris databases or sensors, respectively

Chapter Contents:

• 16.1 Techniques for sensor identification/dataset classification
• 16.1.1 PRNU-based sensor identification (PSI)
• 16.1.2 Iris texture classification (ITC)
• 16.1.2.1 Feature extraction
• 16.1.2.2 Feature encoding
• 16.1.2.3 Classification
• 16.2 Datasets
• 16.3 Experimental setup
• 16.4 Experimental results
• 16.5 Practical discussion
• 16.6 Conclusion
• References

Inspec keywords: image sensors; iris recognition

Other keywords: PSI; image sensors; photo response non-uniformity; PRNU; iris recognition system

Subjects: Image recognition; Computer vision and image processing techniques