An Improvement Technique based on Structural Similarity Thresholding for Digital Watermarking
Amin Banitalebi-Dehkordi, Mehdi Banitalebi-Dehkordi, Jamshid Abouei, Said Nader-Esfahani
AAn Improvement Technique based on StructuralSimilarity Thresholding for Digital Watermarking
Amin Banitalebi-Dehkordi , Mehdi Banitalebi-Dehkordi , Jamshid Abouei , and Said Nader-Esfahani Department of Electrical and Computer Engineering, University of British Columbia, Canada Department of Electrical and Computer Engineering, Ferdowsi University of Mashhad, Iran Department of Electrical and Computer Engineering, Yazd University, Yazd, Iran Department of Electrical and Computer Engineering, University of Tehran, [email protected], [email protected], [email protected], [email protected]
Abstract —Digital watermarking is extensively used in ownership authentication and copyright protection. In thispaper, we propose an efficient thresholding scheme to improve the watermark embedding procedure in an image. For theproposed algorithm, watermark casting is performed separately in each block of an image, and embedding in each blockcontinues until a certain structural similarity threshold is reached. Numerical evaluations demonstrate that our schemeimproves the imperceptibility of the watermark when the capacity remains fix, and at the same time, robustness againstattacks is assured. The proposed method is applicable to most image watermarking algorithms. We verify this issue onwatermarking schemes in Discrete Cosine Transform (DCT), wavelet, and spatial domain.
Keywords—Digital watermarking; structural similarity; wavelet; discrete cosine transform
I. I
NTRODUCTION
The rapid growth of the Internet and multimedia technologies has revealed the need for the copyright protection and the proofof the ownership of digital documents [1]. In particular, with images widely available on the Internet, digital watermarking is acommon way of identifying images and protecting them from unauthorized usage in online advertisements. In this regard, the mostimportant characteristic of watermark casting is its imperceptibility, where a certain degree of the statistical invisibility of theembedded watermark is required. In addition, in most watermarking techniques, it is desirable to embed messages with appropriatelength so that the accurate extraction is assured, and at the same time, the embedded watermark should be perceptually invisibleand robust to common signal processing and intentional attacks. Thus, there exists a trade-off between the imperceptibility,capacity and robustness of the watermarking methods [2].Broadly, watermarking techniques are divided into two categories: i) spatial domain watermarking, and ii) transform domainwatermarking. Spatial domain watermarking approaches, where the mark is directly embedded into each pixel of the host image,benefit from the advantages of a low degree of complexity and delay [3]–[5]. In addition, the temporal/spatial localization of thewatermark in the spatial domain watermarking schemes is automatically achieved. This permits a better characterization of thedistortion introduced by the watermark and reduces the annoying effects. In transform domain watermarking techniques, however,the watermark is inserted into the coefficients of a digital transform of the host asset [6], [7]. For instance, Barni’s work [6] whichwas based on watermarking using the Discrete Cosine Transform (DCT) and the Discrete Fourier Transform (DFT). Of interest isto embed the watermark in the host image using the wavelet domain transform [8]. Authors in [9] utilize a DCT domainwatermarking technique for copyright protection of digital images. They propose a watermarking method with dual detection wherea pseudo-random sequence of real numbers, as a watermark, is embedded in a selected set of DCT coefficients. Jayalakshmi et al.[10] utilize the multi-resolution dependency of the coefficients of contourlet transform for image watermarking. Generally,transform domain watermarks exhibit a higher robustness to attacks than the spatial domain watermarking schemes. Moreimperceptibility of the embedded watermark can be achieved by avoiding the changes into low frequency components of the hostimage. Complexity of this type of embedding might be higher than the complexity of the spatial domain watermarking methods.This issue introduces a trade-off between the robustness and the complexity when using the transform domain watermarkingschemes.ean Squared Error (
MSE ) is one of the quality measures that has been widely utilized in various watermarking schemes due toits low complexity and ease of usage. However, it is shown that visual systems designed based on the usage of the
MSE criterioncannot completely track the human visual perception with a high reliability and accuracy [11]. The above challenge motivates us topropose a new technique for improving on digital watermarking algorithms. Our scheme utilizes the structural similarity (
SSIM ) asa quality criterion instead of common measures such as the
MSE metric. In the proposed scheme, the watermark embedding processis performed on each block of the host image until a certain threshold of the
SSIM in quality is achieved. The insertion of thewatermark sequence in the host image is performed based on an adaptive procedure. The length of the embedded watermark in aloop is increased until a certain threshold of quality is achieved. The stop condition for the watermark insertion process in eachblock of the image is specified by the
SSIM threshold. In one hand, each block for the above process should be big enough to allowthe
SSIM efficiently extract local structural similarities, and on the other hand, blocks with larger sizes may increase thecomplexity.Numerical results show that the block size of 32×32 yields a lower computational complexity in several algorithms. Ouralgorithm is numerically tested for various watermarking schemes in the spatial (pixel) domain as well as DCT and waveletdomains. In our simulations, the conventional watermarking methods in the DCT, wavelet and spatial domains are modified so thatthe capacity is maximized, while the watermarking imperceptibility remains fix. Since our scheme does not change the original waythat the watermark is being embedded, it is applicable to many of the watermarking algorithms.The rest of the paper is organized as follows: In Section II, the structural similarity theory is briefly reviewed. Section IIIdescribes the proposed optimization method. Section IV shows the results of numerical experiments. Finally, in Section V, anoverview of the results is presented. II. S
TRUCTURAL S IMILARITY
Natural images are highly structured, i.e. there exists a high degree of local correlation between the pixels. To extract the localstructural similarity, we follow the primitive formulation of Wang and Bovic in [11], in which the target application is qualityassessment of natural images. The original and distorted images, denoted by x and y respectively, are decomposed to theirluminance, contrast, and structure components as follows: Luminance,2
122 1
Cμμ Cμμx,yl yx yx (1)
Contrast,2
222 2
CCx,yc yx yx (2)
Structure, CCx,ys yx xy (3)where ( µ x ,µ y ) and ( σ x ,σ y ) represent the local sample means and standard deviations of ( x,y ), respectively, and σ xy denotes the samplecross correlation of x and y after subtracting their means. In addition, C i , i = 1, 2, 3, are small positive constants that stabilize eachterm, so that near-zero sample means, variances, or correlations do not lead to numerical instability. The local structural similarity( SSIM ) is formulated as [11]: yxsyxcyxlyxS ,.,.,, (4)The SSIM metric between the original and the reference image pair ( x,y ) is calculated by averaging the local
SSIM, S ( x,y ), overthe image. III. I MPROVEMENT T ECHNIQUE
As mentioned, regardless of the watermarking algorithm that we would like to improve, the enhancement procedure isperformed on each block separately. In each block of the host image, watermark casting is performed via an adaptive procedureexactly the same as the original algorithm that we are modifying, no matter what method is being used for the watermarkembedding. To assure the imperceptibility of the mark, the embedding process of the watermark in each block is divided intoseveral steps, and the embedding continues until a certain threshold on the similarity between the primitive block and the currentatermarked block is achieved. The above process is performed on all the blocks until the watermarked image is obtained (see Fig.1). To verify the performance of this approach and compare with the non-modified algorithms, we embed the sequence generatedby the concatenation of the sequences that have been embedded in all blocks using the proposed method. Thus, the same sequencesare embedded using the original and the improved algorithms. In this way, the capacity remains constant and therefore theimperceptibility can be compared for both methods. Since the type of watermark casting (e.g. in spatial, DCT or wavelet domain) isnot changed, the robustness will remain almost unchanged. To get more insight into the proposed watermarking algorithm in Fig. 1,we summarize the steps of this figure as follows:
Step 1 : Divide the asset image ( A ) into k×k blocks, where a suitable value of k is selected empirically to reduce thecomputational complexity. For our algorithm, an efficient value of k is 32 for the sake of the complexity reduction. Step 2 : Set i = 1, name the i th block of A as “ block ”, and define w block = block. w block will be the i th block after the watermarkinsertion. Step 3 : For a specific watermarking algorithm, embed the watermark sequence in “ block ” to create w block . Step 4 : Compute the
SSIM between “block” and w block . If this similarity is greater than a pre-specified threshold thr2 , then go toStep 3 and embed a new watermark sequence in “block”. The new watermark has a bigger length when compared to the previouswatermark. Watermark is usually a sequence of random bits. The new mark is the concatenation of the previous mark and L newrandom bits, where L is chosen corresponding to the watermark casting algorithm. If the mark is an image, the new mark can be ahigher resolution image. For the Least Significant Bit (LSB) watermarking, the previous mark would be the ( L-1 ) th left-sided bits ofthe mark image, while the new mark would be the L th left-sided bit planes. Step 5 : Repeat Step 4 until the
SSIM becomes less than thr2.
Step 6 : Once the
SSIM value is less than thr2 , the current w block is the ultimate marked block of the watermarked image. Savethis block, set i = i + 1 and reset w block to the next block. Then go to Step 4 and perform the same process for the next block of A .Note that the iterative block embedding procedure is repeated for all of the blocks of the original image. Then, the output of thealgorithm would be the watermarked image W . The proposed algorithm leads to an optimum watermark embedding in the sensethat by adjusting the threshold value, one can directly tune the imperceptibility and the capacity. According to Fig. 1, the structuralsimilarity based method offers an adaptive embedding procedure that embeds the watermark sequence until a certain threshold forthe imperceptibility is achieved. Clearly, less threshold values results in a higher capacity but less invisibility. Another advantage ofthe proposed method is that the scheme is robust against most of the intentional and unintentional attacks. As we will show in Fig. 1. Flowchart of the proposed watermarking algorithm ection IV, for most of the watermark casting algorithms our scheme can extract the true watermark sequence with a higheraccuracy than each original algorithm. IV. N
UMERICAL R ESULTS
This section presents some numerical evaluations for the proposed watermarking algorithm, where we use the Lena image asasset. The proposed method has been examined by modifying four watermarking schemes described briefly in the subsequentsections. Afterwards we will state our innovation. In our simulation, the threshold thr2 for
SSIM is empirically set to 0.8.
A. Pixel Domain Embedding: LSB Watermarking
For the original LSB algorithm, watermark is an image which is embedded in another host picture. For each pixel, the N left-sided bit planes of the mark image are replaced with the N right-sided bits of the asset. In this way, most significant bits (i.e.important information) of the watermark image are replaced with the least significant bits (or details) of the host image. Within anadaptive process, in each time slot, a bit plane is added to the embedding bits of the watermark image block. The embeddingprocess is stopped when the threshold is reached or all bits are embedded. As previously mentioned, the embedding process isperformed adaptively in each block. B. DCT Domain Embedding: Barni’s Method [12]
For the original algorithm described in [12], the watermark is a sequence of random bits placed in an additive form to the DCTcoefficients of the host image as follows:
Mixttt iiLiLiL ...,,2,1, (5)where X = { x , x , …, x M } is the embedding sequence, t j and t′ j are the DCT coefficients of the host and the watermarked images,respectively, and α is a parameter to adjust the watermark power. For a better comparison with our method, the visual maskingeffect (i.e., the extra stage for improving the imperceptibility) is turned off in both the original and the proposed block-wisealgorithm. The same steps are done for the proposed algorithm; the only difference is that the embedding process is performed oneach single block separately. In comparison to the original method, the watermarking sequence length is much smaller for eachblock, but larger for the entire image. The value of α, and the length of the watermark sequence are increased iteratively in eachloop cycle until the similarity threshold is reached ( SSIM between the watermarked and the original blocks meets the threshold).
C. Wavelet Domain Embedding: Ahuja’s Method [13]
For the original algorithm in [13], the watermark casting is performed in the wavelet domain. Authors in [13] utilize the waveletdecomposition property to divide the host image to different frequency resolutions. Similar to the embedding process in the DCTdomain, the watermark sequence is adaptively embedded in the middle frequency wavelet coefficients. For the proposed algorithm,casting of the mark is done in each block separately. The embedding process is done iteratively. Insertion of the watermark in eachof the iterations is performed on 2000 wavelet coefficients more than the previous iteration. The embedding process continues untilthe
SSIM threshold is reached.
D. Wavelet Domain Embedding: CDMA-Based Method [14]
For the original algorithm, the watermark casting is based on the Code Division Multiple Access (CDMA) and multilevelcoding [14]. The bits of the watermark are grouped together and for each sequence a different modulation coefficient is used.Different watermark messages are hidden in the same transform coefficients of the cover image using orthogonal or semi-orthogonal codes [14]. The proposed algorithm is the same as the original one. However, it is performed for each individual blockseparately. Size of the mark increases during each iteration until the
SSIM threshold is reached.Fig. 2 demonstrates the original and the watermarked images for all four algorithms mentioned above. This figure also showsthe effect implied by our method on the images. As observed in Fig. 2, the proposed method improves the imperceptibility propertyof the watermark while the capacity remains constant in both original watermarking method and the proposed approach. Table Iillustrates the statistical invisibility of the watermark. Similarity in terms of the
MSE and the
SSIM between the original and thewatermarked images is given in Table I. It is revealed that the similarity and therefore the imperceptibility are maintained more inthe proposed algorithm than in the original methods. In some cases, the difference is salient. ig. 2. LSB algorithm: original (a), proposed (b), DCT-based algorithm: original (c), proposed (d), wavelet-based algorithm: original (e), proposed(f), CDMA-based algorithm: original (g), proposed (h) he complexity of the proposed method highly depends on the complexity of the original method that our approach ismodifying. As Table I shows, the complexity for the LSB watermarking is highly reduced when it is improved by our method.However, the proposed method increases the complexity for the DCT-based and the wavelet-based watermarking methods. For theCDMA-based scheme, the complexity is not changed. Taking the above considerations into account and this fact that our approachperforms the same trend as the original approach, we conclude that the complexity of the proposed method depends on the originalwatermarking scheme.Fig. 3 demonstrates the response of watermark detectors to attacked watermarked images. After the original image iswatermarked by the proposed methods (i.e. modified version of the four algorithms), the following three types of attacks areapplied to them: Gaussian noise, Low Pass Filtering (LPF), and JPEG compression. Attacks have the same parameters andconditions for all images. The watermark is considered as a sequence of bits, even for the LSB watermarking. 300 randomsequences are generated where three of them are replaced with the extracted sequences from the attacked images. Fig. 3 shows thecorrelations between the true watermark sequence and the above 300 sequences. It is revealed from this figure that the proposedmethod is robust in almost all attacks. This robustness is followed from the robustness of the original algorithms, in particular forthe DCT-based and the wavelet-based algorithms which are robust against attacks.
TABLE I. W
ATERMARK I MPERCEPTIBILITY AND A LGORITHM COMPLEXITY FOR THE ORIGINAL AND PROPOSED METHODS
MSE SSIM
Algorithm Complexity:Simulation Time (Sec)
Original Proposed Original Proposed Original Proposed
LSB 11.28 0.03 0.72 0.98 99 4DCT-based [12] 50.64 0.5 0.5 0.96 2 22Wavelet-based [13] 26.03 16.48 0.75 0.95 4 35CDMA-Wavelet [14] 71.59 9.13 0.42 0.97 228 231
Fig. 3. Watermark detection performance for the proposed method. High bars that stand out in each figure represent Gaussian noise,low pass filtering (LPF), and JPEG compression, from left to right respectively: (a) LSB watermarking, (b) DCT-basedwatermarking, (c) wavelet-based watermarking, and (d) CDMA-based watermarking . Sensitivity of the proposed method to the block sizes
It was mentioned in Section III that the blocks for watermark embedding are of k×k size. We selected k equal to 32 for anefficient trade-off between the performance of the proposed method and computational complexity. In this sub-section, we studythe effect of the block size on the embedding performance as well as the computational complexity. To this end, the performanceand complexity are evaluated at various block sizes. The same parameters as Table I are used so that the SSIM values for theoriginal algorithms are 0.72, 0.5, 0.75, and 0.42 for the LSB, DCT-based [12], Wavelet-based [13], and CDMA-Wavelet [14]methods, respectively. Fig. 4 illustrates the trade-off between the performance, complexity, and the block size. It is observed fromFig. 4 that changes in block size result in smooth variations in the watermark imperceptibility and complexity of differentalgorithms. We chose a block size of 32, however, any other block size can be chosen.
F. Sensitivity of the proposed method to the threshold value
The block embedding threshold value ( thr2 ) specifies the length of the watermark which is embedded in each block. The lowerthe threshold is, the more number of bits will be embedded in each block and therefore a higher capacity is achieved. However, alower threshold results in a more perceptible mark in the host image. To find a suitable threshold value, we sweep the thresholdvalue from 0.6 to 0.9 and evaluated the
SSIM and
MSE values for the marked image compared to the original one. Results areillustrated in Table II. It is observed from Table II that the
SSIM and
MSE values vary smoothly when the threshold changes. Thisshows the robustness of the proposed method to the threshold value. We chose a threshold value of 0.8. However, this value can bechanged according to the desired amounts of imperceptibility and capacity.
Fig. 4. Imperceptibility and complexity for various block sizes: (a) LSB watermarking, (b) DCT-based watermarking, (c) wavelet-basedwatermarking, and (d) CDMA-based watermarking . Validation of the proposed method over a dataset of images
We validate the performance of the proposed method for a dataset of images to verify the generalization of our approach. Fig. 5shows snapshots of the incorporated image set. Using this set of images and the same threshold parameters used in Section IV, theproposed method resulted in an average of 25 %, 44 %, 23 %, and 53 % improvements in visual quality (
SSIM ) for LSB, DCT-based [12], Wavelet-based [13], and CDMA-based [14] watermarking methods, respectively.V. C
ONCLUSION
In this paper, we proposed a new method to improve the watermark embedding procedure. We suggested that many embeddingalgorithms can be performed via each block separately and for all the blocks. This can be performed during an adaptive iterativeprocess. The stop condition for this procedure is a threshold on the structural similarity that is compatible with the human visualsystem. Numerical results showed that the proposed method can improve the imperceptibility of the watermark for variousexamined watermark casting algorithms. It was seen that the capacity remains constant and the complexity varies among differentmarking algorithms. After intentional attacks such as Gaussian noise, low pass filtering, and JPEG compression, the watermarkdetector showed noticeable correlation between the extracted watermarks and the true watermarks which verifies that the proposedalgorithm is robust against these attacks. R
EFERENCES[1] I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich, and T. Kalker, Digital Watermarking and Steganography, Morgan Kaufmann, 2 nd ed., 2008.[2] X. Gong and H.-M. Lu, “Towards fast and robust watermarking scheme for H.264 video,” in Proc. 10 th IEEE International Symposium on Multimedia(ISM) , Dec. 2008, pp. 649–653.[3] S. S. Bedi and S. Verma, “A design of secure watermarking scheme for images in spatial domain,” in Proc. Annual IEEE India Conference , Sept. 2006.[4] H. H. Larijani and G. R. Rad, “A new spatial domain algorithm for gray scale images watermarking,” in Proc. International Conference on Computer andCommunication Engineering (ICCCE) , May 2008, pp. 157–161.[5] V. Licks, F. Ourique, R. Jordan, and F. Perez-Gonzalez, “The effect of the random jitter attack on the bit error rate performance of spatial domain imagewatermarking,” in Proc. International Conference on Image Porcessing (ICIP) , Sept. 2003, pp. 455–458.[6] M. Barni, F. Bartolini, A. De Rosa, and A. Piva, “Capacity of the watermarking channel: how many bits can be hidden within a digital image,” in Proc.Security and Watermarking of Multimedia Contents, Procedeeings of SPIE , Jan. 1999, vol. 3654, pp. 437–448.[7] A. Banitalebi, S. Nader-Esfahani, and A. Nasiri Avanaki, “Robust LSB watermarking optimized for local structural similarity,” in Proc. 19 th IranianConference on Electrical Engineering (ICEE) , May. 2011.
TABLE II. W
ATERMARK I MPERCEPTIBILITY AND A LGORITHM COMPLEXITY FOR VARIOUS THRESHOLD ( THR ) VALUES
Thr2=0.6 Thr2=0.7 Thr2=0.8 Thr2=0.9MSE SSIM MSE SSIM MSE SSIM MSE SSIM
LSB 28.22 0.91 12.39 0.94 0.03 0.98 0.01 1DCT-based [12] 37.78 0.88 14.18 0.93 0.5 0.96 0.13 1Wavelet-based [13] 39.39 0.87 23.96 0.92 16.48 0.95 1.43 0.99CDMA-Wavelet [14] 36.36 0.89 14.07 0.94 9.13 0.97 0.71 1
Fig. 5. Snapshots of the incorporated image dataset
8] Y. Yusof and O. O. Khalifa, “Imperceptibility and robustness analysis of DWT-based digital image watermarking,” in Proc. IEEE International Conferenceon Computer and Communication Engineering (ICCCE) , May 2008, pp. 1325–1330.[9] G. P. Betancourth, A. Haggag, M. Ghoneim, T. Yahagi, and J. Lu, “Robust watermarking in the DCT domain using dual detection,” in Proc. IEEEInternational Symposium on Industrial Electronics , July 2006, pp. 579–584.[10] M. Jayalakshmi, S. N. Merchant, and U. B. Desai, “Significant pixel watermarking in contourlet domain,” in Proc. IET International Conference on VisualInformation Engineering (VIE’06) , Sept. 2006, pp. 416–421.[11] Z. Wang and A. C. Bovik, “Mean squared error: love it or leave it? a new look at signal fidelity measures,”
IEEE Signal Processing Magazine , vol. 26, no. 1,pp. 98–117, Jan. 2009.[12] A. Piva, M. Barni, F. Bartolini, and V. Cappellini, “DCT-based watermark recovering without resorting to the uncorrupted original image,” in Proc.International Conference on Image Processing (ICIP) , Oct. 1997, pp. 520–523.[13] R. Dugad, K. Ratakonda, and N. Ahuja, “A new wavelet based scheme for watermarking images,” in Proc. International Conference on Image Processing(ICIP) , Oct. 1998, pp. 419–423.[14] S. P. Maity and M. K. Kundu, “A blind CDMA image watermarking scheme in wavelet domain,” in Proc. International Conference on Image Processing(ICIP)in Proc. International Conference on Image Processing(ICIP)