Digital image forensics

Jessica Fridrich

IEEE Signal Processing Magazine · 2009 · 327 citations · 17 references

Concepts

TL;DR

Photo‑response nonuniformity (PRNU) is an intrinsic, pixel‑level noise pattern that serves as a sensor fingerprint, enabling device identification, linking, processing‑history recovery, and forgery detection even after lossy compression or filtering. The tutorial demonstrates how to estimate the PRNU fingerprint from camera images and detect it in other images to verify origin and integrity. The authors formulate forensic tasks as two‑channel hypothesis tests and apply a generalized likelihood ratio test to estimate and detect the PRNU fingerprint. Illustrative examples show the effectiveness of the proposed methods.

Abstract

The article explains how photo-response nonuniformity (PRNU) of imaging sensors can be used for a variety of important digital forensic tasks, such as device identification, device linking, recovery of processing history, and detection of digital forgeries. The PRNU is an intrinsic property of all digital imaging sensors due to slight variations among individual pixels in their ability to convert photons to electrons. Consequently, every sensor casts a weak noise-like pattern onto every image it takes. This pattern, which plays the role of a sensor fingerprint, is essentially an unintentional stochastic spread-spectrum watermark that survives processing, such as lossy compression or filtering. This tutorial explains how this fingerprint can be estimated from images taken by the camera and later detected in a given image to establish image origin and integrity. Various forensic tasks are formulated as a two-channel hypothesis testing problem approached using the generalized likelihood ratio test. The performance of the introduced forensic methods is briefly illustrated on examples to give the reader a sense of the performance.

References

17