Iris Recognition Liveness Detection via Corneal Reflection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Iris recognition systems face challenges in distinguishing between live human eyes and spoofing attempts, such as using photographs, due to the time required for data processing and user discomfort from varying light intensities, and existing methods can be easily defeated by placing small illuminators in the iris image.
Innovation Solution
A system that uses a computer screen to reflect an image onto a user's eye, with cameras capturing the reflection to determine if it is consistent with a human eye, employing methods like image magnification, curvature analysis, and illumination pattern recognition to differentiate between live and fake eyes, while also associating face and iris imagery for secure biometric recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If light intensity is varied to detect pupil dilation, then live eye detection is improved, but user discomfort increases
Solution Approach 1:
The system uses periodic modulation of light intensity at specific frequencies to stimulate the pupillary light reflex. By varying light intensity in a controlled periodic manner and detecting the corresponding periodic pupillary response, the system can reliably distinguish live eyes from spoofing attempts without requiring extreme or uncomfortable light levels.
Solution Approach 2:
The system monitors the pupillary response in real-time and uses this feedback to verify liveness. The detected pupillary dilation/constriction patterns provide continuous feedback about the authenticity of the eye being scanned, enabling reliable detection while maintaining comfortable operating light levels.
2Reliability
If multiple images are captured and processed to verify liveness, then spoofing detection is improved, but processing time increases
Solution Approach 1:
The system captures multiple images at different time points during the scanning process, including preliminary images that establish baseline pupillary state. By capturing images sequentially as the scan progresses, the system builds liveness verification data in real-time without requiring post-processing of large datasets, thus reducing overall processing time.
Solution Approach 2:
The system uses rapid sequential imaging to capture pupillary responses at multiple time points quickly. By rushing through the image capture phase with high frame rates and processing only critical frames, the system achieves thorough liveness verification without excessive processing delays.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively deters spoofing attempts by accurately identifying live human eyes and securely associates face and iris imagery, enhancing the reliability and efficiency of biometric recognition without causing user discomfort.
Implementation Method 1
Light from a display screen is reflected by a cornea of a candidate eye
Implementation Method 2
The cornea acts as a convex mirror that magnifies an image reflected therefrom
Data Source
AI summary
A method of biometric recognition is provided. Multiple images of the face or other non-iris image and iris of an individual are acquired. If the multiple images are determined to form an expected sequence of images, the face and iris images are associated together. A single camera preferably acquires both the iris and face images by changing at least one of the zoom, position, or dynamic range of the camera. The dynamic range can be adjusted by at least one of adjusting the gain settings of the camera, adjusting the exposure time, and/or adjusting the illuminator brightness. The expected sequence determination can be made by determining if the accumulated motion vectors of the multiple images is consistent with an expected set of motion vectors and/or ensuring that the iris remains in the field of view of all of the multiple images.