基于MPEG-7和MM混合模型的图像自动标注算法

Algorithm of automatic image annotation based on MPEG-7 and MM mixture model

为了弥补图像低层视觉特征和高层语义之间的"语义鸿沟",改善图像自动标注的性能,提出了基于多媒体描述接口(MPEG-7)和MM(Mixture Model)混合模型的图像标注算法。该算法采用MPEG-7标准推荐的颜色和纹理描述子提取图像的低层视觉特征,通过MM混合模型建立低层特征到高层语义空间的映射,实现了基于图像整体低层特征的多标签图像自动标注。通过在corel图像数据集上的一系列实验测试验证了该方法的可行性和有效性。

To compensate for the ＂semantic gap＂ between low-level visual features of image and high-level semantics and improve the performance of automatic image annotation, a algorithm of image annotation based on Multimedia Description Interface （MPEG-7） is proposed. Low-level visual features of images are extracted according to the color and texture descriptors which are recommended by the MPEG-7 standard, Mapping is setted up from low-level features to high-level semantics space by MM mix- ture model. Image is automaticly annotated with multi-label based on the overall low-level image features. The proposed algo- rithm is demonstrated to be feasible and effective on the corel image datesets.