A superior Support Vector Machine digital watermarking for color image

In this paper, we have proposed an improved & logistic digital watermarking scheme, which collected based on Support Vector Machine (SVM) for color image. The watermark is implanted into the discrete wavelet domain of the original image and extracted by training support vector machine, which have the element of the image. For performance enhancement of support vector machine, we consider the adding of momentum coefficient to reduce the error and enhance the rate of the learning. The watermark can be successfully extracted by training the support vector machine, and the (SVM) watermarking algorithm is good for many kinds of common attacks. The experimental results reveal that the proposed algorithm can achieve the desired result and high corpulence to general image processing technique & geometric distortion.