Image carving with missing headers and missing fragments

Abstract

Although some remarkable advancements have been made in image carving, even in the presence of fragmentation, existing methods are not effective when parts (fragments) of an image are missing. This paper addresses this problem and proposes a PRNU (Photo Response Non-Uniformity)-based image carving method. The proposed technique assumes that the underlying camera fingerprint (camera sensor noise) is available prior to the carving process. Given a large number of image fragments, the camera fingerprint is used to find the position of fragments in a to-be-carved image. Using all known-position-fragments, the number of to-be-carved images is then found. The known-position-fragments and the unknown-position-fragments are placed on these images using two different greedy algorithms. Experiment with 23040 fragments shows that the proposed scheme has a true positive rate of 94.2%.