摘要
针对如何在图像检索系统中客观地表达用户的感知,提出了一种基于粗糙集和遗传算法的相关反馈图像检索方法。采用粗糙集的方法求出图像的最简视觉特征子集,采用交互式遗传算法,对每一次反馈得到的候选图像集进行评价,利用遗传算子从图像库中选出接近用户需求的图像。实验表明该方法能有效的提高检索效率,是一种更为理想的检索方法。
Aimed at presenting objectively the perspectives of users in the system of image retrieval, an image retrieval method of relevant feedback based on rough set and genetic algorithms is proposed. Firstly, the method of rough set is adopted to calculate the simplest visual feature subset of images, then interactive genetic algorithms is used to evaluate the candidate image sets which are gained by every feedback, finally, inheriting operator is used to select precisely the approximately qualified images from the library of images. Additionally, experiments explicitly manifest that the novel approach is a more ideal retrieve method and the speed of image retrieval greatly is enhanced.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第9期2086-2088,共3页
Computer Engineering and Design
关键词
图像检索
相关反馈
视觉特征
粗糙集
约简
交互式遗传算法
image retrieve
relevant feedback
visual feature
rough set
reduction
interactive genetic algorithm
