摘要
图像特征选择是基于内容的图像检索的关键技术之一,ReliefF算法是常用的图像特征选择算法。针对ReliefF特征选择算法的不足,利用信息论中的散度对其进行改进,在相同的时间复杂度下,使得结果的有效性得到一定的改善。提出两步法的特征选择框架,并且实现去除冗余特征的算法,在保证结果有效性的前提下大大降低了时间复杂度。
Image feature selection is one of the key technologies for content-based image retrieval.And ReliefF method is a commonly-used way of image feature selection.The author tried to improve the shortcomings of the ReliefF method by means of Kullback divergence in information theory.As a result,this improvement makes the results of feature selection more efficient in the same time complexity.A new feature selection framework has been proposed which can be achieved in two steps.Using this framework can remove the redundant features effectively and obviously reduce the time complexity on the premise of ensuring the validity of the results.
出处
《丽水学院学报》
2011年第5期34-41,共8页
Journal of Lishui University
关键词
图像检索
特征选择
信息论
image retrieval
feature selecting
information theory
