Forgery-location highway monitoring system based on video watermarking

This paper presents a new content-dependent watermarking algorithm for forgery-detection highway monitor system. The basic scheme is to use weighted partial energy relation between groups of DCT blocks in a macroblock to generate semi-fragile feature codes. Then the codes are sequentially embedded into another macroblock using Differential Energy Watermarking (DEW) by a secret key, which ensures the algorithm robust to ordinary processing. Compared with the conditional DEW algorithm, the watermarking system security is enhanced remarkably and the forgeries can be localized at the macroblock level.