Abstract
With the development of the identification methods, an iris recognition system is expected to be one of the basic elements of modern society, with many application areas such as access control, national ID cards, etc. Any iris recognition system follows four functioning steps: Segmentation, Normalization, Encoding, and Matching. Iris segmentation is a key step in any iris recognition system and it directly affects the accuracy of matching. Several methods have been suggested to assess iris regions with two non-concentric circles. The circle model was investigated to find a tradeoff between modeling complexity, accuracy of the algorithm and computational time. In this paper, we propose an enhancement in the form of a modified Masek approach and a comparative study of the performance of three methods: radial segmentation, Masek approach and ours which are all parts of circle model approaches. Using CASIA Iris Database V3.0, our experimental results reveal that the proposed method provides a high performance in time and accuracy.
| Original language | English |
|---|---|
| Pages (from-to) | 468-477 |
| Number of pages | 10 |
| Journal | International Review on Computers and Software |
| Volume | 9 |
| Issue number | 3 |
| State | Published - 2014 |
| Externally published | Yes |
Keywords
- Biometrics
- Circle model
- Eyelids
- Iris segmentation
- Pupil