|
|
|
| Home > Technology > White Paper > Iris Recognition |
|
|
 |
|
 |
| |
 |
 |
|
 |
| |
 |
|
 Advantages |
A person's iris is fully developed within 18 months after birth, and is protected by eyelashes, eyelids and the retina. Its shape hardly changes so that it has higher consistency compared to other biometric characteristics.
Its higher uniqueness in shape than a face or fingerprints ensures that an authentication system using the iris is immensely reliable.
|
|
| Disadvantages |
- Invasive method.
- Equipment for scanning is costly.
|
 |
| |
 |
| Personal identification using the iris is composed of two parts: obtaining the iris image and recognizing it. Firstly, the system performs the function of obtaining the iris image suitable to iris recognition. The second part is comprised of two stages: extracting the iris area from the image and creating an iris code, and perform a match based on the iris characteristics. |
|
 |
|
| Figure 4.2 above, briefly shows the process of extracting the iris area from the eye image, splitting it, and creating an iris code. |
|
| Iris Input Device |
| Obtaining an eye image from an input device is the first stage of personal identification using the iris. The device is comprised of a camera to capture the image and lighting & image sensors to grab correct iris patterns. In particular, the device is closely related to the system's overall performance. General iris input devices are using multi-wavelength infrared rays to prevent iris patterns from being affected by external light. Unlike face recognition, close-up photographing is required to obtain appropriate resolution of the eye image since an eye is smaller than a face. In the case of close-up photographing, it is difficult to set a clear focus due to low depth of field. To obtain a clear image, the shutter speed of a camera and capture speed of an image sensor should be fast. The infrared rays used to capture the iris image have low intensity, and are known to be non-harmful to humans. The rays can be usefully employed in the iris recognition system after multiple trainings. However, from the users' point of view, having their sensitive eyes scanned through the device could be psychologically offensive. |
|
| Iris Area Extraction |
After capturing the eye image, the iris area should be correctly extracted from it. Detecting the inner boundary of the iris against the pupil and the outer border of the iris against the sclera finishes the process.
Both borders of the iris are determined by approximating them into a circle based on the premise that they are circle-shaped. For the circle detection, extract the inner/outer boundaries of the iris after performing the preprocessing on the eye image, or change "hoop transformation" that is widely used for the detection of the same-shape line. In general, a Circular Edge Detector is commonly used. However, the outer boundary of the iris could have a non-circular shape. At present, various kinds of techniques are being used for more accurate detection of the iris, and are being introduced in dissertations.
|
|
| Extraction and Matching of Iris Features |
| The extraction of iris features means capturing ring-shape patterns around the iris area. Various extraction methods of iris-features were developed. The feature extraction method using the Gabor Wavelet transformation, which is commercialized and recognized for its good performance, is introduced here.
After the iris area is extracted, the area is divided into 8 small ring areas as shown in Figure 4.2. The gray values of the iris patterns are calculated clockwise or counter-clockwiseand the iris feature data in actual use can be obtained through the Gabor Wavelet Transformation of the value.
|
|
|
Iris Recognition
