| 基于协同进化教与学优化算法的图像分割 |
| Co-evolutionary teaching-and-learning optimization algorithm based image segmentation |
| |
| DOI： |
| 中文关键词: 图像分割；二维最大熵；教与学优化算法；协同进化算法 |
| 英文关键词:image segmentation;two-dimensional maximum entropy;teaching-and-learning based optimization(TLBO) algorithm;co-evolutionary algorithm |
| 基金项目:江苏省自然科学基金(BK20170914，BK20160910)、南京邮电大学校级科研基金(NY215047，NY217059，NY214114)和国家自然科学基金(61806100，61701260，61502250)资助项目 |
| 作者 | 单位 | | 孙希霞 | 南京邮电大学 物联网学院，江苏南京210003 | | 白晓东 | 南京邮电大学 通信与信息工程学院，江苏南京210003 | | 许斌 | 南京邮电大学 物联网学院，江苏南京210003 | | 潘甦 | 南京邮电大学 物联网学院，江苏南京210003 |
|
| 摘要点击次数: 1571 |
| 全文下载次数: 863 |
| 中文摘要: |
| 提出了一种基于协同进化教与学优化(Co-evolutionary Teaching-and-Learning based Optimization，CTLBO)算法的二维最大熵多阈值分割方法。首先，给出了二维熵多阈值分割的最优化模型。然后，针对教与学优化（Teaching-and-Learning based Optimization，TLBO）算法存在的早熟收敛和停滞问题，提出了一种CTLBO算法，并将该算法应用于二维熵多阈值分割最优化模型的求解。该算法将整个班级分为多个子班级，每个子班级的学员同时向所有子班级的老师学习，从而提高种群多样性。此外，每隔一定的代数，各子班级的老师组成新的班级进行信息交流，从而提高收敛速度。最后，应用仿真实验对所提方法的有效性和可行性进行了验证。实验结果表明：与基于传统TLBO算法及其相关改进算法、粒子群算法的图像分割方法相比，所提方法具有更好的优化能力和分割性能。 |
| 英文摘要: |
| A co-evolutionary teaching-and-learning based optimization (CTLBO) algorithm based two-dimensional maximum entropy image segmentation method is proposed.Firstly,an optimization model for the two-dimensional maximum entropy segmentation is presented.Then,to overcome the drawbacks of premature convergence and stagnation of the traditional TLBO algorithm,a co-evolutionary teaching-and-learning optimization algorithm is used to slove the optimal thresholds of two-dimensional entropy.The algorithm divides the whole class into several sub-classes,and every student learns from teachers of all the sub-classes,increasing the diversity of the population.Moreover,teachers of all the sub-classes form a class to communicate with each other,increasing the convergence speed.Finally,the feasibility and the effectiveness of the proposed method are tested by simulation experiments.Experimental results demonstrate that the method has optimization and segmentation capabilities compared with the TLBO algorithm and its variants,particle swarm optimization algorithm based segmentation methods. |
| 查看全文 查看/发表评论 下载PDF阅读器 |