Reversible data embedding using difference expansion scheme for watermarking

Rohit Kulkarni,Shriya Tatachar
Abstract:Reversible Data Embedding is a lossless reversible data embedding technique which embeds invisible data known as payload into the input image and uses Difference Expansion (DE) technique. This paper uses a DE based reversible scheme to reduce the amount of auxiliary information embedded and to increase the watermark length. The key concepts discussed in this paper are: Reversible Embedding, Extraction and Restoration of the original dataset with minimal distortion. By observing redundancy, reversibility can be explored. Payload, PSNR and Watermark Length have been used to assess the effectiveness of DE. Four data sets have been used - Lena, Boat, Airplane and APC. For Lena the payload obtained is 31.14 and payload was found to be 0.76 implying that more information can be embedded in an image than in the standard methods.
Computer Science
What problem does this paper attempt to address?