Efficient PRNU Extraction Using Joint Edge-Preserving Filtering for Source Camera Identification and Verification

Abstract

The photo response non-uniformity (PRNU) of a digital image is a kind of sensor pattern noise (SPN) by which one can distinguish the gadget which is utilized to take the image. It has been extensively used for source camera identification and verification in the past decade. The useful data that the SPN passes makes it unique, and thus sensible for perceiving the source camera and recognizing image forgeries. In any case, the PRNU extraction process is unavoidably looked with the contamination of image dependent features and furthermore other non-unique artifacts. In this work, we have presented joint edge preserving filtering technique for Π extraction. The proposed method is computationally efficient and PRNU separated by utilizing the proposed method contains an insignificant measure of scene details of the image. Hence the PRNU obtained from the proposed method is more efficient for source camera identification. Exploratory outcomes on publicly available database by different cameras have demonstrated a significant gain acquired and the predominance of the proposed framework over related best in class frameworks.