一种新的基于SVM和主动学习的图像检索方法

A novel image retrieval method based on SVM and active learning

在基于内容的图像检索中,支持向量机(SVM)能够很好地解决小样本问题,而主动学习算法则可以根据学习进程主动选择最佳的样本进行学习,大幅度缩短训练时间,提高分类算法效率。为使图像检索更加快速、高效,提出一种新的基于SVM和主动学习的图像检索方法。该方法根据SVM构造分类器,通过"V"型删除法快速缩减样本集,同时通过最优选择法从缩减样本集中选取最优的样本作为训练样本,最终构造出不仅信息度大而且冗余度低的最优训练样本集,从而训练出更好的SVM分类器,得到更高的检索效率。实验结果表明,与传统的SVM主动学习的图像检索方法相比,该方法能够较大幅度提高检索性能。

In the content-based image retrieval, Support Vector Machine （SVM） can resolve the problem of small sample size, and the active learning algorithm can select the most optimal samples to learn actively according to the learning process, thus reducing the training time greatly and improving the efficiency of the classification algorithm. In order to obtain the more rapid and efficient image retrieval, a novel image retrieval method based on SVM and active learning is proposed. Firstly, the method constructs the classifier on the basis of SVM. Secondly, the ＂V＂ elimination method is used to reduce the sample sets quickly, and the optimal selection method is applied to select the optimal samples from the reduced sample sets as the training ones. Finally, the optimal training sample set with abundant in- formation and lower redundancy is obtained, so that the better SVM based classifier is constructed and the higher retrieval efficiency is achieved. Experimental results show that, compared with the traditional image retrieval method based on SVM and active learning, the proposed method has better performance and can improve the retrieval performance greatly.