1. An image adaptive, wavelet-based watermarking of digital images
Santa Agreste, Guido Andaloro, Daniela Prestipino, Luigia Puccio
Journal of Computational and Applied Mathematics
Department of Mathematics, University of Messina, Italy
Reporter : 韓培真
2019/10/28

2. Outline
In digital management, multimedia content and data can easily be used in an illegal way being copied, modified and distributed again.
Copyright protection, intellectual and material rights protection for authors, owners, buyers, distributors and the authenticity of content.
WM2.0 for an invisible watermark: private, strong, wavelet-based and developed for digital images protection and authenticity.
WM2.0 works on a dual scheme: watermark embedding and watermark detection.

3. 1. Introduction
In digital management, multimedia content and data can easily be used in an illegal way, this is easier than ever before.
In this paper they put the focus on watermarking techniques and describe the realized wavelet-based algorithm for watermarking of still images.
Another:
1. present an overview about tools and motivation.
2. describe in detail the proposed algorithm.
3. present some experimental results.
4. concludes this work.

4. 1. Introduction Watermark properties Watermark of digital must be：
1. invisible
2. robust to attacks and tamer-resistant
3. low rate of false alarm and high level of secrecy
Attack to watermarked digital images:
1. remove attacks
2. geometrical attacks
3. cryptographic attacks
4. protocol attacks

5. 1. Introduction
The property of wavelet functions is to process data at different scales or resolutions, highlighting both large and small features.
The main advantages of inserting watermarks in the wavelet transform domain:
Space-frequency localization
Multi-resolution representation
Superior HVS modeling
Linear complexity
Adaptivity

6. 1. Introduction
Space-frequency localization :
Space-frequency domain

7. 2. Watermark algorithm
Embed watermark signals into high-frequency sub-bands discrete wavelet transform (DWT) coefficients, according to the HVS directives.
HVS considerations indicate that the eye is less sensitive to noise in those areas of the image where brightness is high or low.

8. 2. Watermark algorithm We denote with Ĩ the watermarked image.
Embedding:
Let C be the matrix associated to block to be watermarked.
Apply to C to obtain the four sub-matrices CKLL,CKHL,CKLH,CKHH of order nk = n / 2k
The DWT coefficients of the
kth decomposition level as elements.
I0HL
I0LH
I0HH
I1LH
I1HH
I1HL
I2HL
I2HH
I2LH
I2LL
DWT with three-level of decomposition.

9. 2. Watermark algorithm
Typically we apply three DWT decomposition levels on a block with associated matrix of order 256.
Watermark embedding is calculated with the formulas:
i, j = 1, 2, ..., nk
Then the value of parameter (i, j) is

10. 2. Watermark algorithm
On watermarked image Ĩ (which can likely be attacked) and original image I the same steps from pre-processing to DWT decomposition are computed
Detection:
Pf is fixed to 10-8 and
If ρ > TP watermark signal is detected.
otherwise, watermark signal is not detected.

11. 3. Attacks and experimental results
Experimentation results have shown than WM2.0 is robust against attacks of geometric, filters and StirMark, with a ration of more than 88%.
It has a low probability of false positive alarm.
P.18 shows the values of p and Tp for different choices of Pf from to
Elaboration time depends on blocks subdivision of the image and thus DWT levels decomposition.

12. 3. Attacks and experimental results
(a) Lena: value (HSV) plane of the original image.
(b) Lena: difference between the original (value) matrix and the watermarked matrix.

13. 3. Attacks and experimental results
(c) Lena: watermark matrix in related level decomposition.
(d) Lena: DWT coefficients before and after watermark embedded.

14. 3. Attacks and experimental results
(a) Lena: watermarked image without attacks, plot of ρ and Tρ. (b) Lena: watermarked image with StirMark attack, plot of ρ and Tρ. (c) Lena: watermarked image with geometric attack (cutout), plot of ρ and Tρ.

15. 4. Conclusions
The watermark embedded has a high level of robustness against geometric and image processing attacks and a low rate of false alarm.
Using DWT on HSV color model in this algorithm for robustness, invisibility and signal watermark distribution over image.