The Community for Technology Leaders
Image and Graphics, International Conference on (2013)
Qingdao, China China
July 26, 2013 to July 28, 2013
pp: 106-111
The credibility of digital images is decreasing with the fast development of easy-to-use image manipulation tools such as Photoshop and Picasso. Although good forged images are deceptive to human eyes, the manipulation process often leaves some statistical traces in the forged images. Very often, the forged images are rotated partly or as a whole. If we can identify image rotation or even determine image rotation angle, the authenticity of images could be verified. In this paper, we propose an image rotation angle estimation algorithm based on properties of 2-D DFT of pixel variance. This algorithm is a blind one because it works in absence of any prior information such as digital watermark. We show the efficacy of this approach in our experiment.
Mathematical model, Discrete Fourier transforms, Equations, Estimation, Interpolation, Algorithm design and analysis, Reactive power

R. Qian, W. Li, N. Yu and B. Liu, "High Precision Image Rotation Angle Estimation with Periodicity of Pixel Variance," 2013 Seventh International Conference on Image and Graphics (ICIG), Qingdao, China, 2013, pp. 106-111.
1041 ms
(Ver 3.3 (11022016))