PROJECT TITLE :

Preprocessing Reference Sensor Pattern Noise via Spectrum Equalization

ABSTRACT:

Although sensor pattern noise (SPN) has been proved to be an efficient means to uniquely establish digital cameras, some non-distinctive artifacts, shared among cameras endure the identical or similar in-camera processing procedures, typically give rise to false identifications. So, it's desirable and necessary to suppress these unwanted artifacts thus as to boost the accuracy and reliability. During this paper, we have a tendency to propose a unique preprocessing approach for attenuating the influence of the non-distinctive artifacts on the reference SPN to cut back the false identification rate. Specifically, we equalize the magnitude spectrum of the reference SPN through detecting and suppressing the peaks consistent with the native characteristics, aiming at removing the interfering periodic artifacts. Combined with six SPN extractions or enhancement methods, our proposed spectrum equalization algorithm is evaluated on the Dresden image database furthermore our own database, and compared with the state-of-the-art preprocessing schemes. The experimental results indicate that the proposed procedure outperforms, or at least performs comparable with, the prevailing ways in terms of the general receiver operating characteristic curves and kappa statistic computed from a confusion matrix, and tends to be a lot of proof against JPEG compression for medium and little image blocks.