| 引用本文: | 邹德旋,王鑫,段纳.一种基于修正差分进化的虹膜定位算法(英文)[J].控制理论与应用,2013,30(9):1194~1200.[点击复制] |
| ZOU De-xuan,WANG Xin,DUAN Na.Iris location algorithm based on modified differential evolution algorithm[J].Control Theory and Technology,2013,30(9):1194~1200.[点击复制] |
|
| 一种基于修正差分进化的虹膜定位算法(英文) |
| Iris location algorithm based on modified differential evolution algorithm |
| 摘要点击 2721 全文点击 1620 投稿时间：2012-09-29 修订日期：2013-05-25 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2013.12223 |
| 2013,30(9):1194-1200 |
| 中文关键词 修正差分进化 虹膜定位 混沌序列 变异操作 中心解 去噪 |
| 英文关键词 modified differential evolution iris location chaotic sequence mutation operation center solution denoising |
| 基金项目 This work was supported by the National Natural Science Foundation of China (No.61104222), Natural Science Fundamental Research Project of Jiangsu Colleges and Universities (No.11KJB510026), Science Fundamental Research Project of Jiangsu Normal University (No.9212812101). |
|
| 中文摘要 |
| 提出一种用于虹膜定位的差分进化算法(modified differential evolution, MDE). MDE和原始差分进化算法(differential evolution, DE)主要有3点不同: 第一, MDE采用了基于混沌序列的尺度因子和基于均匀分布的交叉率, 这有助于提高候选解的多样性; 第二, MDE使用中心解来修正最差解的变异操作, 这有助于提高候选解的质量; 第三, MDE使用最好解来帮助受困解摆脱局部最优点. 在搜索边缘前, 两种有效的去噪方法被用来减少虹膜图像中噪声的影响. 去噪后, 再使用MDE和其他4种方法来进行虹膜定位. 在中科院(Chinese Academy of Sciences Institute of Automation, CASIA)眼图数据库中选择200幅来自不同个体的虹膜图像来验证和比较MDE及其他4种方法的效率. 实验结果表明, 与其他4种方法相比, MDE使用更少的执行时间来定位瞳孔边缘和虹膜边缘. |
| 英文摘要 |
| A modified differential evolution (MDE) algorithm is proposed for iris location. The MDE and the original differential evolution algorithm are different from three aspects: First, MDE adopts the scale factor based on chaotic sequences and the crossover rate based on uniform distribution, which is helpful to improve the diversity of candidate solutions. Second, MDE utilizes the center solution to modify the mutation operation of the worst solution, which is beneficial in improving the quality of candidate solutions. Third, MDE uses the best solution to help the trapped solutions to escape from local optima. Before searching boundaries, we use two kinds of efficient denoising methods to reduce the effects of noises on iris edge images. After denoising, the proposed MDE and the other four methods are applied to iris location. Some 200 iris images of different individuals are chosen from the Chinese Academy of Sciences Institute of Automation (CASIA) eye image database in investigating and comparing the efficiency of MDE with the other four methods. Experimental results show that MDE consumes less execution time to locate pupil and iris boundaries in comparison with the other four methods. |
|
|
|
|
|