An efficient encoding based watermarking technique for tamper detection and localization

In the present era of preposterous melioration, the security of the image data has got prioritized. Although a large number of methods have been developed for security and authenticity verification of the images, still the prevalent data breaches raise the concern of the development of novel methods for such a purpose. Digital image watermarking is viewed as a potential approach for ensuring the integrity and authenticity of image data. To address the important authentication issue, this research proposes an efficient encoding-based watermarking scheme for tamper detection and localization. Firstly, the cover image is divided into 4 x 4 non-overlapping blocks and the Arithmetic Average (AA) of each block is computed. This value is represented by an 8-bit Vector (BV). This BV is enhanced to a 16-bit vector by appending 8-bits of an encrypted watermark. The resultant Enhanced Bit Vector (EBV) is embedded into 4 x 4 blocks of the image. For better imperceptibility, Huffman encoding is applied to the watermark while the security of the watermark is taken care of using DNA encryption. The EBV serves the purpose of tamper detection while BV serves to localize the tamper. The experimental results reveal that the proposed technique has a lower computing complexity in addition to tamper detection and localization capability. Furthermore, the suggested approach provides an average PSNR of 51.41 dB, revealing that the watermarked images exhibit high visual quality for a payload of 512,416 bits i.e., 1.95 bpp.