Image Watermarking Algorithm Based On Principle Components Analysis

As the developing rapidly network technique, various digital multimedia information is widely spread via the network, including texts, images, audio frequency, video frequency, etc. Meantime, as the information on the internet can be copied and modified conveniently, tort on the internet is becoming increasingly serious. As a result, how to protect copyright of works has already been extensive attention. In this case, digital watermarking technology has become a research hotspot, and the image digital watermarking technology is the focus of the study, which has made a lot of research results at home and abroad.PCA is the typical method of signal processing technology. It makes dimension reduction of original complex data, and picks up significant and irrelevant characteristic variables. These characteristic variables, which are the largest part of image energy and have better robustness, can confront noises or other disturbances. In consideration of the above characteristics of the Principle Components, in this essay, we will combine Principal Components Analysis technology and Digital Watermarking technology and bring forward Image Digital Watermarking Algorithm based on PCA. In the algorithm,first,we selects the principle components which can best represent carrier image characteristics through demonstrating experimental analysis, and then inserts watermarks into the principle components. Experimental results show that the algorithm has strong anti-attack ability and good stability.The main work of this paper is as follows:(1)To make an overview of the current principles of digital watermarking technology features, typical algorithms, attack methods and the evaluation of watermarking algorithms and bring about problems and solutions in designing watermarking algorithm.(2) To state in detail the basic principle of PCA, implementation methods and application, and analyze the characteristics of principal components.(3) To present image digital watermarking algorithm based on principal component analysis (PCA).First, we block the vector image and make analysis of the principle components of each part, and then pick up the principal components whose energy is more than 85% as the embedding position of the watermark.At last, we test watermark whether there with the relevant. (4) To propose the imperceptible and robust of the algorithm.Through a large of experiments, the results show that the algorithm has well imperceptible by comparing with DCT and DWT algorithm. In addition, it has strong robustness to some common attacks, such as adding noise, cropping, low pass filtering, median filtering and JPEG compression.