A Secure Semi-Fragile Watermarking Scheme for Authentication and Recovery of Images based on Wavelet Transform Rafiullah Chamlawi, Asifullah Khan, Adnan Idris, and Zahid Munir Abstract—Authentication of multimedia contents has gained much attention in recent times. In this paper, we propose a secure semi-fragile watermarking, with a choice of two watermarks to be embedded. This technique operates in integer wavelet domain and makes use of semi fragile watermarks for achieving better robustness. A self-recovering algorithm is employed, that hides the image digest into some Wavelet subbands to detect possible malevolent object manipulation undergone by the image (object replacing and/or deletion). The Semi-fragility makes the scheme tolerant for JPEG lossy compression as low as quality of 70%, and locate the tempered area accurately. In addition, the system ensures more security because the embedded watermarks are protected with private keys. The computational complexity is reduced using parameterized integer wavelet transform. Experimental results show that the proposed scheme guarantees the safety of watermark, image recovery and location of the tempered area accurately. Keywords—Integer Wavelet Transform (IWT), Discrete Cosine Transform (DCT), JPEG Compression, Authentication and Self- Recovery. I. INTRODUCTION HE ease, by which digital multimedia data can be manipulated, has always raised many concerns about the possibility to reliably trust their content. Digital data authentication is thus one of the most important and investigated security applications. In our proposed approach, the image authentication and recovery is based on a comprehensive technique that operates with the computing of two watermarks [1], an image digest and a binary image. The image digest is computed through a properly modified version of JPEG coding operating at very high compression ratio on original image [2]. Thus image digest is a compressed version of the image itself and it helps in obtaining an estimate of the original contents. The modification is introduced in the digest to make it insensitive to global, innocuous manipulations. The other watermark, binary signature (second watermark) is processed with a private key to ensure security [3]. The scheme is flexible enough with the choice of users, either to embed image Rafiullah Chamlawi, and Asifullah Khan are with Pakistan Institute of Engineering and Applied Sciences (PIEAS), Islamabad, Pakistan (e-mail: chamlawi@gmail.com, asif_jg@yahoo.com). Adnan Idris is with AJK University, Rawalakot, Azad jamu and Kashmir, Pakistan (e-mail: adnaidris@gmail.com). Zahid Munir is with Electronic Division PINSTECH, Nilore, Islamabad, Pakistan (e-mail: mzahid@pinstech.org.pk). digest, binary image or both. Embedding binary image can help in accurately detecting manipulations made in image, but it cannot ensure recovery of an estimated image. Similarly embedding image digest can retrieve the estimated image but leaves the users to judge the authenticity by themselves. Thus, embedding image digest as well as binary image can lead to both authentication and recovery. For the reason we use image digest as a compressed version of the original image, our technique can also be referred as a self- recovery technique. The scheme use the parameterize integer wavelet transform which is the fast approach of Discrete Wavelet Transform. Based on the idea, [4] proposed for the first time to use the parameterized wavelet transform. However, his scheme is still based on conventional DWT. Lifting scheme is an effective method to improve the processing speed of DWT. Integer wavelet transform allows to construct lossless wavelet transforms. By lifting scheme, we can construct integer wavelet transform. In this paper, we will address the secure semi-fragile watermarking for image authentic-cation and recovery based on integer wavelet transform with parameters. In current communication, we discuss the watermarks generation, embedding and extraction in section 2, section 3 explains the temper detection. We report experimental results in section 4 and conclusion are made in section 5. II. WATERMARK GENERATION Our scheme is based on embedding of two watermarks. We proceed for the watermarks generation in this section. A. Binary Image Preprocessing A binary signature is preprocessed before embedding as a watermark. Let W be a binary signature of size M×N and PN be a pseudorandom matrix of same size generated by a secret key. The binary signature W and pseudorandom matrix PN are represented as; (, ) (1 ,1 ) W wi j i M j N = ≤≤ ≤ ≤ (1) where { } (, ) 0,1 wij ∈ (, ) (1 ,1 ) n RandomMatrix pij i M j N = ≤≤ ≤ ≤ (2) where { } (, ) 0,1 n p ij ∈ We adopt the formula (3) to get the ultimate watermark 1 W : T World Academy of Science, Engineering and Technology International Journal of Computer and Information Engineering Vol:2, No:11, 2008 3866 International Scholarly and Scientific Research & Innovation 2(11) 2008 scholar.waset.org/1307-6892/3308 International Science Index, Computer and Information Engineering Vol:2, No:11, 2008 waset.org/Publication/3308