摘要
在研究图像检索基本理论基础上,提出基于主色提取和主色集扩充的图像检索算法,并运用基于BP网络的相关反馈方法提高算法性能,通过开发图像检索系统并进行检索实验,验证了算法的检索性能及BP网络相关反馈算法的有效性。
Basing on studying the basic theory of image retrieval, this paper propose an image retrieval algorithm based on dominant color extraction and dominant color set expansion, and improve performance of the algorithm by using BP network for relevance feedback. Through retrieval experiments with the image retrieval system, it has proved the effectiveness of the algorithm with relevance feedback based on BP network.
基金
辽宁省高校重点实验室基金(2008S115)
关键词
图像检索
神经网络
相关反馈
主色特征
image retrieval
neural network
relevance feedback
dominant color characteristics
