**3. Experimental results**

In this section, we show some experimental results to demonstrate the effectiveness and success of our digital watermarking techniques. The standard 512 × 512 grayscale image "pepper" is used as host image, as shown in Fig. 7. The 32 × 32-pixels binary image is used as the watermark image, as shown in Fig. 8.

We applied the peak-signal to noise rate (PSNR) given in (4) to measure the image quality of an attacked image and accuracy rate AR given in (5) to evaluate the robustness of a copyright protection scheme for a specific attack.

A DFT-DWT Domain Invisible Blind Watermarking

K=50

Fig. 9. (i) Watermarked image, (ii) Extracted watermark image

K=30

has to be performed.

Table 1.

**4. Conclusion** 

binary image as watermarked image.

Techniques for Copyright Protection of Digital Images 525

Exhaustive testing against signal processing operation, Geometric distortion , collusion still

**Gain Factor(K) Execution Time AR PSNR**  K=10 176.6796 89.4532 60.3456 K=20 175.6307 93.0640 55.2389 K=30 173.3867 91.3765 54.7829 K=40 175.3912 93.5621 54.6047

The need for digital watermarking on electronic distribution of copyright material is becoming more prevalent. In this paper an overview of the digital watermarking techniques are given and a blind invisible watermarking technique for grayscale images based on DWT and DFT is presented. The algorithm use 512\*512 gray images as a host image and 32\*32

Firstly, two level wavelet decomposition is implemented on the host image. Then, the middle frequency components are extracted and divided in to several blocks of size 4\*4 and DFT is implemented on them. Finally, two pseudo random sequences are created and embedded to blocks which have implemented DFT according to whether the corresponding

The original image is not required while extracting the watermark Instead, correlations among each block and two sequences are respectively calculated. Watermark is recovered on foundation of the relative magnitude of correlation between the corresponding block and one sequence or the other. The idea of applying two transform is based on the fact that

position is 0 or 1 in the watermark matrix which has been implemented.

Fig. 7. Original image of pepper

Fig. 8. Watermark image
