Paper Info
Title
A Robust Image Tampering Detection Method Based on Maximum Entropy Criteria
Abstract
This paper proposes a novel image watermarking method based on local energy and maximum entropy aiming to improve the robustness. First, the image feature distribution is extracted by employing the local energy model and then it is transformed as a digital watermark by employing a Discrete Cosine Transform (DCT). An offset image is thus obtained according to the difference between the extracted digital watermarking and the feature distribution of the watermarked image. The entropy of the pixel value distribution is computed first. The Lorenz curve is used to measure the polarization degree of the pixel value distribution. In the pixel location distribution flow, the maximum entropy criteria is applied in segmenting the offset image into potentially tampered regions and unchanged regions. All-connected graph and 2-D Gaussian probability are utilized to obtain the probability distribution of the pixel location. Finally, the factitious tampering probability value of a pending detected image is computed through combining the weighting factors of pixel value and pixel location distribution. Experimental results show that the proposed method is more robust against the commonly used image processing operations, such as Gaussian noise, impulse noise, etc. Simultaneously, the proposed method achieves high sensitivity against factitious tampering.
Year
DOI
Venue
2015
10.3390/e17127854
ENTROPY
Keywords
Field
DocType
image tampering detection,watermarking,maximum entropy,offset image
Mathematical optimization,Feature detection (computer vision),Pattern recognition,Free boundary condition,Image processing,Probability distribution,Artificial intelligence,Pixel,Principle of maximum entropy,Morphological gradient,Gaussian noise,Mathematics
Journal
Volume
Issue
Citations 
17
12
0
PageRank 
References 
Authors
0.34
7
3
Name
Order
Citations
PageRank
Bo Zhao100.68
Guihe Qin2239.00
Pingping Liu3174.99