median filter is not only used to achieve blurring but also it is used to
reduce the noise. In image enhancement applications, median filter modifies the
center pixel value of the window with the middle value of the sorted pixel
values. The proposed method is examined in opposition to median filtering
attacks with different window size of 2 x 2, 3 x 3. The extracted watermark
images and its corresponding median filtering modified image are shown in Fig.
NOISE ADDITION ATTACK
(ii) Speckle Noise
Speckle noise is multiplicative
noise exists inherently like a granular noise. The variance of single pixel and
local area are equal that is centered on that pixel. This method is tested
against speckle noise under different noise variances are depicted in Table3. It
can be easily understood that extracted watermark image quality is improved
with decrement of noise variant. NCC values of extracted watermark image is inversely
proportional to the noise variance and BER is directly proportional to the
noise variance of 0.004 is shown in Fig. .
(iii) Gaussian Noise
Gaussian noise is one of the
statistical noise processing operations. Variance with zero mean varies the
amount of noise. It is used to prove the robustness of the proposed method by the addition of statistical noise
to encrypted watermarked image as shown in Table 3. Encrypted watermarked image
and extracted watermark image as described in Fig. .
value of an image is not differing too much by applying small rotation but the
encrypted watermarked image can be affected. Hence the robustness against
rotation attack is tested with 22.50 angle shown in Fig. .
Imperceptibility compatibility of the proposed method is represented by PSNR
values in the Table .
is used to focus on necessitate part of an image and the remaining part of an
image is eliminated. Cropping of an image make changes to the aspect ratio and
improve image framing. Here the robustness of the proposed technique is
experimented against cropping operation by the window size shown in Fig. . The
proposed method are demonstrated in Table . by the PSNR, BER, SSIM and NCC.
(vi) Brightness and contrast
contrast attack is one of the common attack on digital multimedia images. Here the
brightness and contrast of the image is
increased and decreased as shown in Fig. . Then the robustness of the image is
tabulated in Table using PSNR, BER, SSIM and NCC.
(vii) JPEG compression
compression is used for manipulation or compression of an image. By changing
the quality factor values for JPEG compression is examined. Extracted watermark
images NCC and BER values from the compressed encrypted watermarked image are
depicted in Table and in Fig..