Global Adaptive Optimization Parameters For Robust Pupil Location
Yang Wang, Xiaoyi Lu, Wenjun Zhou · 2021 17th International Conference on Computational Intelligence and Security (CIS) · 2021
Most existing pupil positioning methods use adaptive threshold in image binarizing, however, this approach could barely completely separate the pupil from the eye, indicating difficulty in screening the pupil contour and affecting the accuracy of pupil location. To overcome this weakness, we present a novel pupil positioning method based on global adaptive optimization parameter, which involves setting a three-step threshold (TST). First, TST binarizes image through searching for the most adaptive threshold corresponding to each eye image. Then a separate pupillary region is obtained. Finally, set the judgment conditions of the pupil connected area, and the connected area of pupillary region is identified by the morphological characteristics of pupil. TST has been tested on the Institute of Automation of the Chinese Academy of Sciences (CASIA-IrisV1) dataset with the recognition accuracy at 99.7%. The experimental results have been compared with other state-of-the-art methods, in which our proposed method show a better performance.