Name
Title | ||
|---|---|---|
Identification of Natural Images and Computer Generated Graphics Using Multi-fractal Differences of PRNU. |
Abstract | ||
|---|---|---|
Comparing with computer generated graphics, natural images have higher self-similar and have more delicate and complex texture. Thus, the distribution of multi-fractal dimensions and singular index of natural images general have large variation range. Based on this, multi-fractal spectrum features of photo response non-uniformity noise PRNU are used for the identification of natural images and computer generated graphics. 9 dimensions of texture features including the square of the maximum difference in fractal dimension SMDF, the square of the maximum difference in the singularity indices SMS and the variance of fractal dimensions VF are extracted from LL, LH, HL sub-bands of PRNU after wavelet decomposition. The identification is accomplished by using LIBSVM classifier. Experimental results and analysis indicate that it can obtain an average identification accuracy of 99.69ï¾ź%, and it is robust against resizing, JPEG compression, rotation and additive noise. |
Year | DOI | Venue |
|---|---|---|
2015 | 10.1007/978-3-319-27122-4_15 | ICA3PP |
Field | DocType | Citations |
Graphics,Computer vision,Wavelet decomposition,Fractal dimension,Computer science,Fractal,Singularity,Artificial intelligence,Jpeg compression,Classifier (linguistics) | Conference | 1 |
PageRank | References | Authors |
0.35 | 10 | 3 |