交互式差分演化策略图像空间滤波方法

Image Spatial Filtering Method Based on Interactive Differential Evolution Strategy

针对传统图像空间滤波方法适应性差和算法缺少交互等问题,提出基于交互式差分演化策略的图像空间滤波方法.该方法由结合差分演化算法及演化策略思想设计的交互式差分演化策略和演化函数矩阵构成.同时,交互式差分演化策略通过引入人的主观评价控制演化过程,使滤波结果更符合人的视觉需求,同时也使该算法具有更好的鲁棒性和全局优化能力.针对交互式中人工评价产生的噪声问题,该方法采用演化函数矩阵建立适应值近似评价模型,实现了主观适应值的自动评价,有效减少了人的评价次数,从而减少了评价噪声.实验结果表明该方法具有更强的全局优化能力和更好的滤波效果,明显优于传统图像空间滤波方法和基于演化策略的图像空间滤波方法.

For the problems of traditional image spatial filtering methods,such as poor adaptability and lack of interaction,etc.,this paper presents a new image spatial filtering method based on interactive differential evolution strategy. This method consists of interactive differential evolution strategy w hich combines the concepts of differential evolution w ith of evolution strategy and the evolution function matrix. M eanw hile,interactive differential evolution strategy introduces human＇ s subjective evaluation to control the processes of evolution. It makes filtering results more in line w ith the needs of the human visual and also makes interactive differential evolution strategy has better robustness and global optimization capability. For the noise problem arising from the interactive evaluation,the proposed method uses the evolution function matrix to establish the fitness approximation evaluation model to achieve the automatic evaluation of the subjective fitness. It effectively reduces the number of human＇s evaluation,thereby reducing the evaluation noise. The experimental results show that the proposed method has stronger global optimization capability and better filtering effect and is significantly better than the traditional image spatial filtering methods and the image spatial filtering method based on evolutionary strategy.