Contourlet Based Adaptive Watermarking for Color Images

Abstract

SUMMARYThis paper proposes a contourlet based adaptive watermarking for color images (CAWCI). A color image with RGB space is firstly converted to its YCbCr space equivalent; a luminance (Y) image and two chrominance (Cb and Cr) images are subsequently transformed into contourlet domain respectively; the watermark is embedded into the contourlet coefficients of the largest detail subbands of three images lastly. On the one hand, the embedded watermark is imperceptible because contrast sensitivity function and watermark visual mask are adopted in our CAWCI. On the other hand, the embedded watermark is very robust due to the spread specialty of Laplacian pyramid (LP) in contourlet transform. The corresponding watermarking detection algorithm is proposed to decide whether the watermark is present or not by exploiting the unique transform structure of LP. Experimental results show the validity of CAWCI in terms of both watermarking invisibility and watermarking robustness. key words: contourlet, image watermarking, information security, wavelet

A dual adaptive watermarking scheme in contourlet domain for DICOM images

“…The embedding capacity, PSNR, and SSIM (in ROI) for both 512 × 512 CT and MRI images in our method are respectively 0.0077 bpp, 60dB and 0.997, while these metrics for 440 × 888 CT images in [ 8 ] are 0.5168 bpp, 50.7038 dB and 0.9876 and for 512 × 512 MRI images are 0.5238 bpp, 41.2629 dB and 0.9558. For [ 44 ], the results of SSIM are not available but for 512 × 512 CT and MRI images PSNRs in dB are respectively 46.47 and 46.37. In this method the total number of embedded bits (for both modalities) are 5358 while in our method, this number is 2010.…”

“…In addition, we achieve the embed process in the lowpass subband to improve the robustness and invisibility. However, the invisibility may be enhanced if the embed process is done in the detail subbands [ 44 ].…”