基于Zernike矩与SIFT特征的商标检索算法  

Trademark retrieval based on Zernike Moment and SIFT features

作　　者：卜宪宪[1] 韩仲志[1] 邓立苗[1] PU Xianxian;HAN Zhongzhi;DENG Limiao(College of Science and Information Science,Qingdao Agricultural University,Qingdao Shandong 266109,China)

机构地区：[1]青岛农业大学理学与信息科学学院,山东青岛266109

出　　处：《太赫兹科学与电子信息学报》2020年第6期1065-1072,共8页Journal of Terahertz Science and Electronic Information Technology

基　　金：国家自然科学基金资助项目(31201133);山东省自然科学基金资助项目(ZR2017MC041);山东省重点研发计划资助项目(2017GNC10111)。

摘　　要：为降低商标检索算法的误检率,提出一种结合Zernike矩(ZM)和尺度不变特征变换(SIFT)的商标检索算法,该算法由离线数据库构建和在线检索组成。分别从查询图像中提取ZM和SIFT特征;根据查询图像的特征集与数据库中存储的图像的特征集之间的ZM特征进行相似度度量,形成候选商标集;最后,利用SIFT特征对查询图像与候选图像精准检测,对相似距离进行排序,将结果返回给用户。实验结果表明:与当前流行的商标检索算法相比,该算法具备更好的检索性能,在缩放、平移、模糊、透视、斜切、扭曲等变换干扰下,仍呈现出更理想的Precision-Recall曲线以及F值。In order to reduce the false detection rate of trademark retrieval algorithm,a trademark retrieval algorithm combining Zernike Moment(ZM)and Scale Invariant Feature Transform(SIFT)is proposed.The proposed algorithm consists of offline database construction and online retrieval.Firstly,Zernike Moment and SIFT features are extracted from the query image.Then,according to the similarity measure between the feature set of the query image and the Zernike moment feature set of the images stored in the database,the candidate trademark set is formed.Finally,the SIFT feature is utilized to detect the query image and candidate image accurately,and the similarity distances are sorted,and the result is returned to the user.The experimental results show that,compared with the current popular trademark retrieval,the proposed algorithm has excellent retrieval performance and is effective in scaling,translation,blurring,perspective,skew,distortion and other transformation forms.The scheme shows good Precision-Recall curve and F value,good real-time performance,strong robustness,and has a certain practical role in trademark registration,auditing and protection.

关 键 词：商标检索 ZERNIKE矩 SIFT特征 Euclidean度量 候选商标 相似距离 

分 类 号：TN957[电子电信—信号与信息处理]