摘要
现有基于传感器光响应非均匀性模式噪声的源相机辨识方法多采用蛮力搜索(即穷举)策略,当相机指纹数量较多且指纹长度较长时,效率较低.文中提出一种基于模式噪声分量符号信息的快速源相机辨识方法.即利用模式噪声中的大幅值分量构造相机的精简指纹,采用分离链接哈希表结构存储参考指纹信息,按照测试指纹和参考指纹之间的同符号分量个数对候选参考指纹排序,按序进行相关性检测.实验表明,文中方法大大减少了相关性运算的次数.
In the existing source camera identification methods based on the PRNU(Photo Response Non-Uniformity) pattern noise of sensors,brute-force search schemes,namely exhaustive schemes,are commonly employed,which may result in low identification efficiency when there exist many long camera fingerprints.In order to solve this problem,a fast source camera identification algorithm is proposed based on the sign information of the components of pattern noise.In this method,compact camera fingerprints are constructed by using the large-magnitude components of pattern noise,and a separate-chaining Hash table is employed to store the information of reference fingerprints.Then,candidate reference fingerprints are permuted based on the number of the components with the same sign between the test and the reference fingerprints.According to the order of reference fingerprints,the correlation detection can be successfully implemented.Experimental results show that the proposed method greatly lightens the computational burden.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2011年第12期1-6,共6页
Journal of South China University of Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60772115)
关键词
源相机辨识
相机指纹
相关性检测
快速搜索
分离链接哈希表
source camera identification
camera fingerprint
correlation detection
fast search
separate-chaining Hash table
