A Secure and Localizing Watermarking Technique for Image Authentication

In this paper, a new block-based image-dependent watermarking technique is proposed. The proposed technique utilizes the correlation coefficient statistic to produce a short and unique representation (also known as hashed values or string-sequences) of the image data. These string-sequences are signed by an error-correcting-code signature scheme, which produces short and secure signatures. The image's least significant bits are utilized to embed these signatures. The used signature scheme requires string-sequences to be decodable syndromes. While the proposed correlation coefficient statistic function produces decodable syndrome string-sequences, most of the existing cryptographic hash functions do not. The results show that the proposed technique has an excellent localization property, where the resolution of the tracked tampered areas can be as small as 9x9 pixel blocks. In addition, the produced watermark has multi-level sensitivity that makes this technique well suited to the region-ofsecurity-important approach, which increases the overall system performance.