摘要
针对MPEG-7单一描述符特征检索效率低的问题,提出了主颜色和边缘直方图特征融合的检索算法。同时利用检索目标在图像中分布的特点,利用分块加权法来提高对检索主目标的检索准确度。实验表明,该方法在查准率和查全率均优于单一特征的检索算法,检索精度得到了提高。
For MPEG-7 low efficiency problems in image retrieval by using a single feature retrieval , the paper presents a new content-based image retrieval method on main color and edge histogram feature fusion. Retrieval accuracy can be improved by retrieval target characteristics of the distribution and block-weighted. Experiments show that the method in the precision rate and recall rate is better than a single feature of the retrieval algorithms, retrieval accuracy is improved.
出处
《电脑与信息技术》
2011年第6期12-15,共4页
Computer and Information Technology
