Improving the Watermarking Technique to Generate Blind Watermark by Using PCA & GLCM Algorithm

Main Article Content

Pratibha Kanawat, Ms. Shalini

Abstract

For making sure that the multimedia information is not accessed or modified by unauthorized users, several digital techniques have been proposed as per the growth of internet applications. However, the most commonly used technique is the watermarking technique. The spatial domain method and frequency domain method are the two broader categorizations of several watermarking techniques proposed over the time. The lower order bits of cover image are improved for embedding a watermark through the spatial domain technique. Minimizing the complexity and including minimum computational values are the major benefits achieved through this technique. However, in the presence of particular security attacks, the robustness of this technique is very high. Further, the techniques that use some invertible transformations such as Discrete Cosine Transform (DCT) are known as the frequency domain transform techniques. The image is hosted by applying Discrete Fourier transforms (DFT) and Discrete Wavelet Transform (DWT) techniques. The coefficient value of these transforms is modified as per the watermark for embedding the watermark within the image easily. Further, on the original image, the inverse transform is applied. The complexity of these techniques is very high. Also, the computational power required here is high. The security attacks are provided with more reverts through these methods. GLCM (Gray Level Co Occurrence Matrix) technique is better approach compare with other approach. In this work, GLCM (Gray Level Co Occurrence Matrix) and PCA (Principal Component Analysis) algorithms are used to improve the work capability of the neural networks by using watermarking techniques. PCA selects the extracted images and GLCM is used to choose the features extracted from the original image. The output of the PCA algorithm is defined by using scaling factor which is further used in the implementation. In this work, the proposed algorithm performs well in terms of PSNR (Peak Signal to Noise Ratio), MSE (Mean Squared Error), and Correlation Coefficient values. The proposed methods values are better from the previous work.

Article Details

Section
Articles