Research Of Digital Datermarking Technology Based On Feature Points

Digital products are vulnerable to accidental or malicious destruction in storage,transmission, processing, and even illegal theft and tampering with the product, damage the legitimate interests of the owners. Digital watermarking technology is an effective means to resolve copyright disputes and protect intellectual property rights. Watermark imperceptibility ensures the raw data availability, also has the robustness to ensure watermark to withstand the attack continues to maintain a certain stability, can be detected watermark, to determine whether the raw data is malicious tampering, maintain its true and reliable.In this paper based on the "electronic eyes" practical background, analyzes the characteristics of the image and video, this paper presents a digital watermarking algorithm based on feature points. Study on the extraction method of Sift algorithm, Harris algorithm and improved Harris algorithm of feature space, calculation of three kinds of algorithm’s time complexity. Among them, the time complexity of Sift algorithm is the highest, and the time complexity of Harris algorithm is the lowest, however, its robustness is poor. By deducing the Harris algorithm formula of response function, eliminated an unknown, saved a lot of time of the experiment, this paper proposed an image watermarking algorithm based on improved Harris feature point. Experimental results show that the proposed algorithm improves the anti geometric attack, anti noise ability, especially the anti noise ability of the original Harris algorithm, and it is widely used in the practical application. This paper studies the 3D-Harris algorithm, Gabor filter and improved Gabor filter. By analyzing the temporal causality of the video, the common filter is discarded, and the causal filter is designed to improve the detection method of space-time feature points. On this basis, a video watermarking algorithm is proposed, which combines the space time feature points and DCT domain. The experimental results show that the proposed algorithm can effectively resist various attacks in time domain and space domain.