中文摘要：

为充分发挥外观设计专利数据不同模态特征的优点,提高检索准确率,提出一种融合文本特征与底层视觉特征的多模态图像检索算法。提出有效内容提取算法提取图像有效内容,对图像有效内容提取底层视觉特征,融合图像文本特征与视觉特征实现多模态检索。实验结果表明,对外观设计专利图像分别检索相似和同类图像时,该检索算法检索效果优于已有单模态检索方法。

英文摘要：

To give full paly to the advantages of different modal features,improve the retrieval accuracy,multi-modal image retrieval algorithm integrating texture features and the underlying visual features was proposed.Efficient contents extraction algorithm was proposed to extract efficient image content.Underlying visual features were extracted from image efficient content,multi-modal image retrieval was done by the fusion of text and visual features.Experimental results show that the proposed retrieval method has better performance compared with the existing single-modal method on retrieving similar images and same class images.