The Research For Geometrical Invariant Digital Watermarking Techniques Based On SIFT

In the recent years, digital watermarking technology has become an effective way to protect copyright and ensure information security. The existing digital watermarking algorithm can get a better effect to image compression, filtering, signal processing attacks, but doesn't resistant to rotation, scaling, translation, shearing and other geometric attacks. The second-generation digital watermarking which using feature points is a research hotspot in recent years.This paper analyzes the existing method of extracting feature points and proposes the research program which using anti-rotation SIFT to extract the steady feature points. The feature points extracted by anti-rotation SIFT are extreme points which found in the scale space and invariant to extraction of position, scale, rotation. In this way , we can ensure the stability of feature points and reduce the dimension of feature descriptors and enhance the speed of matching feature points. Then, we apply the theory of scale-invariant to digital watermarking, and proposed the program which is resistant to geometric attacks based on anti-rotation SIFT. In this program, we extract feature points by anti-rotation SIFT, and select the stable feature points Watermarking is embedded in local area, using odd-even quantization in spatial domain after select the stable feature points. And we regulate the rotation and scaling by SIFT center angle and SIFT center distance. Finally, we validate the presented algorithms through a series of experiments. The results show that anti-rotation SIFT can reduce the matching dimension of feature points, improve the matching rate. And the proposed scheme achieves good image quality and it is robust to geometric attacks as well as traditional signal operations.