Passive image tamper detection based on Fast Retina Key Point Descriptor

摘要

Images are major resource for information owing to their wide use and availability in the internet. Ensuring trustworthiness for image content is a primary objective of image forensic discovery. Copy-move is one of the main and frequently used tampering practices for reproducing or concealing precise area of an image. Most of existing methods deploy block matching technique to match and locate the tampered region which leads high computational complexity. This article introduces a robust technique to discover copy-move tampering with Fast Retina Key Point Descriptor (FREAK). Suspected image is preprocessed and FREAK feature descriptors around each Harris corner points were discovered. Detected FREAK features are mapped by means of KD-Tree algorithm. Experimental outcomes reveals that proposed technique efficiently detects copy-move tampered region even it is rotated, multi-pasted, blurred or brightness adjusted.