摘要
低层特征的选择与提取是自动图像分类的基础,一方面,所选择的图像特征应能代表各种不同的图像属性,利于不同类别图像之间的区分;另一方面,为了提高后续模型的计算效率,需要减少噪声特征、冗余特征。提出了一种基于特征加权的自动图像分类方法。该方法根据图像低层特征分布的离散程度来衡量特征相对于类别的重要性,增加相关度高的特征的权重,降低相关度低的特征权重,从而避免后续模型被弱相关或不相关的特征所支配。所提的特征加权算法主要考察的是特征相对某个具体类别的重要程度,可以为每个类别选择出适合自身的特征权重。然后,将加权特征嵌入到支持向量机算法中用于自动图像分类,在Corel图像数据集上的实验结果表明,基于特征加权的自动图像分类算法可以有效地提高图像分类的准确性。
The low-level feature selection and extraction are the basic problems for automatic image annotation.On the one hand,the selected features must represent the various characters of the images and be beneficial to classify the images.On the other hand,we should limit the dimension of the feature vector and reduce the redundancy of the features to save the posterior computing energy.In this paper,an automatic image classification based on the weighted feature is proposed.This method determines relevant features based on their statistical distribution and assigns greater weight to relevant features as compared to less relevant features,which will avoid the classification model being dominated by weak relevant or irrelevant features.Then we combine the weighted feature algorithm with the support vector machine （SVM） algorithm to realize the image annotation.Experimental results on the Corel image set show that the weighted feature image classification method can effectively improve the performance of classification.
出处
《微型电脑应用》
2014年第1期13-17,共5页
Microcomputer Applications
关键词
自动图像分类
特征加权
支持向量机
Automatic Image Classification
Weighted Feature
Support Vector Machine （SVM）
