Iris Localization Algorithm based on Effective Area

Iris localization is the most crucial part of iris processing because its accuracy can directly affect biometric identification in subsequent steps. Yet, quality images may be sharply degraded due to interference from eyelashes and reflections during image acquisition, which adversely. To solve problem, an algorithm based on effective area proposed. First, YOLOv4 used crop obtain area, beneficial improving localization. Furthermore, a method remove reflective noise proposed, effectively avoid problem process inner boundary determination. Finally, aiming at edge deviation caused by eyelashes, outer adjustment The experimental results show that proposed achieves good performance both outperforms other state-of-the-art methods.