Robust-Fragile Watermarking Model: A User Based Approach for Digital Images

  • Samriddhi Sharma, Saransh Goyal, Vartika Sharma

Abstract

Various techniques of watermarking have been proposed for authentication and protection of Intellectual Property Rights of one’s media before. Different applications demand use of different watermarking techniques. In this research work, we have proposed one such watermarking model which assists the user in making the choice between two secure methods of watermarking for images, fragile and robust method, based on the user’s need. The fragile method employs the hybrid Discrete Wavelet Transform (DWT)- Discrete Cosine Transform (DCT) scheme. The robust scheme in the proposed model is an integration of DWT, Singular Value Decomposition (SVD) and DCT. Type-II DWT with Haar wavelet is applied on the cover image which results in four frequency bands. We also introduced an encoding technique, using an embedding factor, which couples the diagonal matrix of host object with the diagonal matrix of the watermark. Inverse DCT and DWT give us the watermarked image which can be securely transferred over the communication channel. On the receiver end, a similar decoding technique is put in place to extract back the original watermark. Steganography techniques have been implemented using MATLAB as a simulator. The experimental outcomes of our integrated robust and fragile methods are weighted against different conventional watermarking techniques and final outcome reflects that the proposed model demonstrates improved values of PSNR, MSE, SSIM and NCC, which implies that it embeds the watermark more efficiently and also that the extracted watermark has better quality.

Published
2020-06-01
How to Cite
Samriddhi Sharma, Saransh Goyal, Vartika Sharma. (2020). Robust-Fragile Watermarking Model: A User Based Approach for Digital Images. International Journal of Advanced Science and Technology, 29(08), 4042-4054. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/26220
Section
Articles