Performance Evaluation of Visible Digital Image Watermarking Algorithms

Multimedia data has become much easier to access due to the rapid growth of the Internet. While this is usually considered an improvement in everyday lifestyle, it also makes unauthorized copying and distributing of multimedia images, data much casier, therefore presenting a challenge in the field of copyright protection i.e. digital watermarking. Which involves insertion of copyright information/images into the cover image, has been proposed to solve the problem. To embed the watermark robustly and imperceptibly, watermark bits are added to the significant coefficients of each sub band selected by considering the human visual system (HVS) characteristics. This paper, discuss the features and concepts pertaining to the three popular visible watermarking algorithms and analyze them with metrics such as MSE and PSNR i.e. a) wavelet based watermarking technique b) Tian's difference expansion technique c) LSB-prediction error expansion technique. We have discussed wavelet based algorithm, a classical and best suitable for compression but suffer from very low PSNR value due to multi resolution nature and dependability on the quality of the input cover image and the embedding image. Tian's difference-expansion technique is a high-capacity, reversible method for data embedding. However, the method suffers from undesirable distortion at low Sr. Asst. Professor, C.V.R College of Engineering, Hyderabad, A.P, India.embedding capacities and lack of capacity control due to the need for embedding a location map. It is recommended to adopt new reversible data- embedding technique called Least Significant Bit prediction-error expansion which better exploits the correlation inherent in the neighborhood of a pixel and provides an effective method for data embedding than the difference-expansion scheme. The experimental results for many Benchmark/Popular test inages show that LSB prediction-error expansion doubles the maximum embedding capacity and involves significantly less computational complexity with higher PSNR value but also imposes various stages of difficulty. to an intruder who intends to extract the original image from the encrypted sub-blocks when compared to other algorithms.