Information Security through Semi-Fragile Watermarking

Protection of multimedia content has recently become an important issue because of the transition from analog to digital technologies. Unlike the encryption which codes the information into unreadable form steganography and data hiding allow readability of the information along with digital stamps embedded into it preserving the ownership of it. In this paper, we decompose gray scale still image into frequency domain using discrete wavelet transform (DWT). Our generalization includes all four bands (LL, HL, LH, and HH) up to second level. We propose a new algorithm which embed multiple, multilevel and multi-energy watermarks into the image for multi-purposes. We further tested the proposed algorithm against various attacks. The technique is robust against incidental attacks and fragile against intentional attacks in the communication channel. We found that embedding the watermark in lower frequencies is robust to a group of attacks, and embedding the watermark in higher frequencies is robust to another set of attacks. The use of cryptographic keys adds additional security to the system.