CAJ | 학술논문

随着云计算蓬勃发展，越来越多的隐私数据在云计算平台存储和处理。保护隐私数据往往需要信号处理前先进行加密操作，而传统的信号处理方法前提是知晓信号内容，对于密文信号无法处理，这对基于明文的传统图像检索服务提出了重大挑战。为了在确保查询图像和图像数据库隐私的情况下仍能有效地进行图像检索，本文提出了一种新的有效加密域图像检索方法。该方法利用基于空间高低通滤波器的LBP算子与基于HSV彩色空间的颜色直方图分别提取图像的纹理特征和颜色特征。同时采用非对称保持内积大小的加密方法对特征加以加密并外包给云端服务器。云端服务器利用授权的用户所提交的加密图像特征与数据库里的加密特征索引进行比较匹配，并返回查询结果。实验结果表明该方法在保证查询和图像数据库的隐私情况下，能取得很好的检索性能。
수착운계산봉발발전，월래월다적은사수거재운계산평태존저화처리。보호은사수거왕왕수요신호처리전선진행가밀조작，이전통적신호처리방법전제시지효신호내용，대우밀문신호무법처리，저대기우명문적전통도상검색복무제출료중대도전。위료재학보사순도상화도상수거고은사적정황하잉능유효지진행도상검색，본문제출료일충신적유효가밀역도상검색방법。해방법이용기우공간고저통려파기적LBP산자여기우HSV채색공간적안색직방도분별제취도상적문리특정화안색특정。동시채용비대칭보지내적대소적가밀방법대특정가이가밀병외포급운단복무기。운단복무기이용수권적용호소제교적가밀도상특정여수거고리적가밀특정색인진행비교필배，병반회사순결과。실험결과표명해방법재보증사순화도상수거고적은사정황하，능취득흔호적검색성능。
With the rapid development of cloud computing,more and more privacy data are stored and processed in the cloud. In order to protect privacy data, encryption usually needs to be implemented before data processing. However, in the traditional signal processing, encrypted signal can not be processed and pose challenges to traditional image retrieval services based on plaintext search. In order to preserve data privacy of query image and image database, this paper proposes an effective encrypted image retrieval algorithm. Texture features and color features are extracted using LBP based on high-low pass filter and color histogram based on HSV color space respectively. These two features are encrypted by an asymmetric scalar-preserving encryption method and outsourced to cloud server, which allows an authorized user to securely search over it with a corresponding valid encrypted feature and retrieve images of interest. Experimental results show that the proposed method can obtain very good retrieval performance while ensuring privacy of query and image database.