Lossless Recovery Process of Original Images by Removing

Chapter 6 Removable Lossless Visible Watermarking in

6.3 Proposed Method for Removable Lossless Visible Watermarking in

6.3.2 Lossless Recovery Process of Original Images by Removing

In this section, we describe the proposed lossless recovery process of the original

images by removing visible watermarks. First, a binary watermark is extracted from the watermarked tetromino-based mosaic image by the lossless data recovery process, which has been described in Section 5.2.3. Then, a raw palette is constructed by counting the colors of the watermarked image. By using the raw palette, a generation of two color palettes, P_s and P_r, are performed. At last, we replace the colors of the black embedded pixels Ib to recover the original tetromino-based mosaic image by using the mapping between P_s and P_r. The detailed algorithm of the proposed lossless recovery process is described as follows.

Algorithm 6.5. Lossless recovery of tetromino-based mosaic image images.

Input: a watermarked tetromino-based mosaic image I.

Output: a recovered tetromino-based mosaic image I′.

Steps:

Step 1. Extract a binary watermark W from I by the lossless data recovery process (Algorithm 5.2)

Step 2. For each tetromino t in I, perform the following steps.

2.1. If t is in the watermark area, add all pixels in t into a set of black embedded pixels I_b.

Step 6. Randomly pair every two pixels of Ib together using K to form |I_b|/2 pairs of pixels, A.

Step 7. For each pair a in A, swap the two colors of a.

Step 8. For each pixel x in Ib, remove W from I to generate a recovered palette image I′ by the following steps.

8.1. Obtain a color Cn of x.

8.2. Search Cn in P_r to obtain a color index n of C_n. 8.3. Obtain a color Cn′ from Ps by the index n.

8.4. Replace Cn of x with C_n′.

6.4 Experimental Results

Figures 6.3 and 6.4 show some experimental results of applying the proposed removable visible watermarking method for a tetromino-based mosaic image. Figure 6.2 shows a binary watermark used in the experiments. Figures 6.3(b) and 6.4(b) are two watermarked tetromino-based mosaic images by the proposed watermark embedding process (Algorithm 6.4). Figures 6.3(c) and 6.4(c) are two recovered images extracted from Figures 6.3(b) and 6.4(b), respectively, with correct keys by the proposed lossless recovery process (Algorithm 6.5). Figures 6.3(d) and 6.4(d) show the two recovered images extracted from Figures 6.3(b) and 6.4(b), respectively, with wrong keys. As seen in these experimental results, the embedded watermark area looks visually different from the pixels adjacent to them. In addition, the watermark can be removed losslessly with the right key

Figure 6.2 A binary watermark image of size 256×256.

(a)

Figure 6.3 An experimental result. (a) A tetromino-based mosaic image. (b) A watermarked image created from (a) with the watermark shown in Figure 6.2 embedded. (c) A recovered image created with a right key. (d) A recovered image created with a wrong key.

(b)

(c)

(d)

(a)

Figure 6.4 An experimental result. (a) A tetromino-based mosaic image. (b) A watermarked image created from (a) with the watermark shown in Figure 6.2 embedded. (c) A recovered image created with a right key. (d) A recovered image created with a wrong key.

(b)

(c)

(d)

6.5 Discussions and Summary

In this chapter, a removable lossless visible watermarking method based on Chen and Tsai’s method [32] has been proposed. Using the properties of tetromino-based mosaic images, we can embed a binary watermark into a cover image by replacing the colors according to a mapping between two color palettes. In the mean time, the binary watermark is also embedded in the watermarked image for the use in the later recovery process. As a result, it can be used in the copyright protection of tetromino-based mosaic images. Then, a process of lossless recovery of the original tetromino-based mosaic images which are used to embed watermarks is performed by an inverse process of the watermarking method. Authorized users can remove the embedded watermark losslessly from the watermarked image without any additional information. Some experimental results were generated by the proposed watermarking algorithms to prove the feasibility of the proposed method.

Chapter 7 Image Steganography in

Tetromino-based Mosaic Images by Watermarking

7.1 Idea of Proposed Method

Steganography is a science of communicating secret data between senders and receivers. By the senders, the secret data are embedded into the files of certain forms, called cover media, to create stego-media as camouflages. Then, the stego-media are transmitted through public channels, such as the Internet. The receiver acquires the stego-media from the channels and extracts the secret data from the stego-media. The advantage of steganography over cryptography is that the behavior of the communication is evident but the content of the communication is camouflaged.

Invaders who intend to steal the secret may consider that these stego-media are ordinary files and ignore them easily.

As the name suggests, image steganography means that images are used both as the secret data and as the cover media, to perform steganography. In this chapter, we will describe the proposed image steganographic method by a watermarking technique based on the method which has been described in Chapter 6.

7.2 Proposed Method for Image

Steganography in Tetromino-based Mosaic Images by Watermarking

7.2.1 Image Embedding Process

In this section, we describe the proposed image embedding process for image steganography by watermarking. At the beginning, an adapted weighted Euclidean distance derived from Formula 6.2 is proposed. A traditional Euclidean distance between two colors C1 and C2 and the adapted one for a color Ci in a color palette are defined in Formulas (7.1) and (7.2) , respectively, below:

By these two formulas, a process for generation of two color palettes, Ps and Pr, and a one-to-one mapping between them are described in Algorithm 7.1. Next, we embed a secret image into the watermark area by replacing the colors of the watermark area according to the mapping between P_s and P_r. The detailed algorithm of the proposed image embedding process by watermarking is described in Algorithm 7.2.

Algorithm 7.1. Process for generation of two color palettes.

Input: a cover palette Pc, a secret palette Pt, a set of colors C = {C0, C1, …,Cm} of Pc, a set of occurrences O = {O0, O1, …,Om} of C, a set of colors C′ = {C⁰′,

C₁′, …,Cn′} of Pt, and a set of occurrences O′= {O0′, O1′, …,On′} of C′.

Output: a sorted color palette Ps and a rearranged color palette Pr. Steps: weighted Euclidean distances D_w of all the colors of the sorted palette are computed using the colors C of the cover palette Pc and the occurrences O of C. The reason for doing this is that we want to hide the secret image in the watermark area by replacing its colors in such a way to make the watermark area look more obvious. The detail of the proposed image embedding process is described in Algorithm 7.2.

Algorithm 7.2. Image embedding process for image steganography.

Input: a tetromino-based mosaic image I, a binary watermark W, a secret image S, and a secret key K.

Output: a watermarked stego-tetromino-based mosaic image I′.

Steps:

Step 1. For each pixel p in I, perform the following steps.

1.1. If p is in the watermark area, add p into a set of black embedded pixels Ib. 1.2. If p is in the non-watermark area, add p into a set of white embedded pixels

I_w. rearranged color palette P_r by Algorithm 7.1.

Step 7. Obtain a width w and a height h from the secret image S.

Step 8. Divide Ib into two areas, and name them a secret area I_s and a non-secret area

Step 11. For each pair a in A, swap the two colors of a to generate a watermarked image I′.

Step 12. For each tetromino in I_w, hide W, w, h, and P_s in I_w by the data hiding process using small tetromino color shiftings described by Algorithm 5.1.

7.2.2 Image Extraction Process

The proposed image extraction process is an inverse process of the image embedding process. The detail of the proposed image extraction process is described as follows.

Algorithm 7.3. Image extraction process for image steganography.

Input: a watermarked stego-tetromino-based mosaic image I.

Output: a secret image S.

Steps:

Step 1. Extract a binary watermark W, a width w, a height h, and a sorted palette Ps

from I by the lossless data recovery process described by Algorithm 5.2.

Step 2. For each tetromino t in I, perform the following steps.

2.1. If t is in the watermark area, add all pixels in t into a set of black embedded pixels Ib.

Step 5. Use Pc, C, C′, O and O′ to generate a rearranged color palette Pr by Algorithm

Figure 7.1 shows a binary watermark and two secret images which were used in the experiments we conducted to implement the proposed method described previously. Figures 7.2 and 7.3 show some experimental results of applying Algorithms 7.2 and 7.3. Figures 7.2(b) and 7.3(b) are two watermarked stego-tetromino-based mosaic images yielded by the proposed image embedding process (Algorithm 7.2). Figures 7.2(c) and 7.3(c) are two secret images extracted from Figures 7.2(b) and 7.3(b), respectively, with correct keys by the proposed lossless recovery process (Algorithm 7.3). Figures 7.2(d) and 7.3(d) show the two secret images extracted from Figures 7.2(b) and 7.3(b), respectively, with wrong keys.

As seen in these experimental results, secret images can be embedded imperceptibly in the watermark areas, and hidden data can be extracted correctly, by the proposed method.

7.4 Discussions and Summary

In this chapter, an image steganographic method based on the proposed watermarking method has been proposed. Using the properties of tetromino-based mosaic images, we can hide a secret image into a cover image by replacing the colors of the watermark area according to a mapping between two color palettes. In the mean time, the binary watermark and a color palette are also embedded in the watermarked image for the use of the later image extraction process. As a result, it can be used for the covert communication via tetromino-based mosaic images. Then, an image extraction process of the secret images is performed by an inverse process of the image embedding method. Users can embed some secret images in tetromino-based mosaic images by watermarking. Some experimental results were generated by the proposed image steganographic method to prove the viability of the proposed method.

(a) (b) (c)

Figure 7.1 Input images for the proposed method. (a) A binary watermark. (b) A secret image. (c) A secret image.

(a)

Figure 7.2 An experimental result. (a) A tetromino-based mosaic image. (b) A watermarked stego-image created from (a) with a secret image shown in Figure 7.1(b) embedded. (c) The secret image extracted with a right key. (d) The secret image extracted with a wrong key.

(b)

(a)

Figure 7.3 Another experimental result. (a) A tetromino-based mosaic image. (b) A watermarked stego-image created from (a) with a secret image shown in Figure 7.1(c) embedded. (c) The secret image extracted with a right key. (d) The secret image extracted with a wrong key.

(b)

Chapter 8 Conclusions and Suggestions for Future Works

8.1 Conclusions

In this study, we have proposed methods for art image creation and data hiding in them. Users can easily generate art images and embed secret data in them simultaneously. The embedded data can be a secret message, a watermark, or a secret image. We can apply these methods for various information hiding applications, such as covert communication, copyright protection, etc.

A novel type of mosaic image which is called tetromino-based mosaic image is created in this study. In addition, some information hiding techniques are also developed for covert communication, watermarking, and image steganography applications.

In the creation process of tetromino-based mosaic images, we found out all the possible combinations of tetrominoes by the proposed tree enumerating algorithm.

Then, a tetromino database was constructed with different colors and tetromino combinations for the creation of tetromino-based mosaic images. Finally, we also proposed a border enhancement process for improving visual effects in tetromino-based mosaic images.

For data hiding in tetromino-based mosaic images, we utilized two features of the mosaic images to embed data. One is the combinations of tetrominoes, and

another is the color values of tetrominoes. We have proposed a data hiding technique by using distinct combinations of tetrominoes. To improve the visual quality of stego-tetromino-based mosaic images, an edge fitting method was used to enhance edge effects by adjusting combinations of tetrominoes using the information obtained by edge detection. In addition, the feature of the colors of tetrominoes was also used to hide data by increasing or decreasing color values of tetrominoes slightly. Using a lossless data recovery process which combines the functions of data extraction and lossless cover image recovery, we can extract data from stego-images and recover the original tetromino-based mosaic image without any distortion.

For watermarking in tetromino-based mosaic images, we have proposed a removable lossless watermarking method by replacing the colors according to a mapping between two color palettes. Based on this invertible watermarking method, we have proposed an image steganographic method by replacing the colors of the watermark area according to the proposed mapping between two color palettes.

According to our research, we can claim that if there is an image feature of an art image which can be modified and detected, a corresponding data hiding technique can be developed. And the three applications of information hiding, copyright protection, covert communication, and image authentication, can be achieved.

8.2 Suggestions for Future Works

In this study, we have proposed a method for the creation of tetromino-based mosaic images, as well as two data hiding techniques for tetromino-based mosaic images. Furthermore, a removable lossless watermarking method and an image steganographic method by watermarking have also been proposed. However, there are still some interesting topics which are worth further study. They are listed as follows.

For tetromino-based mosaic image creation:

1. Creating combinations of tetrominoes with different sizes and tiling styles to make the tetromino-based mosaic images have various visual effects.

2. Enhancing the performance of creation of tetromino-based mosaic images.

3. Applying tetromino-based mosaic images on characters or geometric totems.

4. Producing mosaic images with other shapes based on computer graphics.

5. Enhancing the performance of the tree enumerating algorithm for finding all possible tetromino combinations by applying A^* search algorithm.

For data hiding in tetromino-based mosaic images:

1. Combining two proposed data hiding methods and applying them on mosaic images to enlarge the embedding capacity.

2. Using more tetromino features such as position, tetromino border, etc. for embedding more data into tetromino-based mosaic images.

3. Extending the proposed methods to survive some attacks such as scaling, rotation, JPEG compression, etc.

For watermarking and image steganography in tetromino-based mosaic images:

1. Enhancing the robustness of the proposed watermarking method.

2. Creating different types of watermark effects in tetromino-based mosaic images.

3. Enlarging the data capacity of the proposed image steganographic method by utilizing other features of tetromino-based mosaic images.

References

[1] R. Breukelaar, E. D. Demaine, S. Hohenberger, H. J. Hoogeboom, W. A. Kosters, and David Liben-Nowell, “Tetris is hard, even to approximate,” International Journal of Computational Geometry and Applications, Vol. 14, pp. 41-68, 2004.

[2] S. Battiato, G. Di Blasi, G. M. Farinella and G. Gallo, “Digital mosaic frameworks - an overview,” Proceedings of Eurographics Italian Chapter, Catania, Italy, February 2006.

[3] C. Close, and J. Yau, Recent Paintings. New York: Pace Wildenstein, 2005.

[4] L. Harmon, “The recognition of faces,” Science American, Vol. 229, No. 5, pp.

70-82, 1973.

[5] S. Dali, The Salvador Dali Museum Collection. Bulfinch Press, 1996.

[6] P. Haeberli, “Paint by numbers: abstract image representations,” Proceedings of SIGGRAPH 90, pp. 207-214, Dallas, USA, 1990.

[7] A. Hausner, “Simulating decorative mosaics,” Proceedings of SIGGRAPH 01, pp.

573-580, Los Angeles, USA, August 2001.

[8] K. Hoff, J. Keyser, M. Lin, D. Manocha and T. Culver, “Fast computation of generalized voronoi diagrams using graphics hardware,” Proceedings of SIGGRAPH 99, pp. 277-286, Los Angeles, USA, August 1999.

[9] S. Lloyd, “Least square quantization in PCM,” IEEE Transactions on Information Theory 28, pp. 129-137, 1982.

[10] Y. Dobashi, T. Haga, H. Johan and T. Nishita, “A method for creating mosaic images using voronoi diagrams,” Proceedings of Eurographics 02, pp. 341–348, Saarbrucken, Germany, September 2002.

[11] G. M. Faustino and L. H. De Figueiredo, “Simple adaptive mosaic effects,”

Proceedings of SIBGRAPI 2005, Natal, Brazil, August 2005.

[12] G. D. Blasi and G. Gallo, “Artificial mosaics,” The Visual Computer, Vo. 21, pp.

373-383, 2005.

[13] G. Elber and G. Wolberg, “Rendering traditional mosaics,” The Visual Computer, Vol. 19, pp. 67-78, 2003.

[14] R. Silvers and M. Hawley, Photomosaics. Henry Holt and Company, 1997.

[15] J. Kim and F. Pellacini, “Jigsaw image mosaics,” Proceedings of SIGGRAPH 02, pp. 657-664, San Antonio, USA, July 2002.

[16] Y. J. Gi, Y. S. Park, S. H. Seo1 and K. H. Yoon, “Mosaic rendering using colored paper,” The 7th International Symposium on Virtual Reality, Archaeology and Cultural Heritage(VAST), pp. 25-30, Baltimore, USA, October 2006.

[17] N. F. Johnson and S Jajodia, “Exploring steganography: seeing the unseen,”

IEEE Computer, pp. 26-34, 1998.

[18] Y. C. Chiu and W. H. Tsai, “Copyright protection by watermarking for color images against rotation and scaling attacks using peak detection and synchronization in discrete fourier transform domain,” Proceedings of 3rd Workshop on Digital Archives Technologies, pp. 207-213, Taipei, Taiwan, August 2004.

[19] Z. Ni, Y. Q. Shi, N. Ansari, and W. Su, “Reversible data hiding,” Proceedings.

IEEE Int. On Cirsuits and Systems, Vol. 2, pp. 912-915, Bangkok, Thailand, May 2003.

[20] Y. J. Cheng, C. Y. Yin, D. C. Wu, and W. H. Tsai, “Copyright protection and authentication of image information in digital libraries,” Proceedings of 2001 Conference on Trends and Perspectives in Digital libraries and Digital Museums, Vol. 3, pp. 1-11, Hsinchu, Taiwan, 2001.

[21] W. L. Lin and W. H. Tsai, “Data hiding in image mosaics by visible boundary regions and its copyright protection application against print-and-scan attacks,”

Proceedings of International Computer Symposium 2004, Taipei, Taiwan, December 2004.

[22] S. C. Hung, D. C. Wu, and W. H. Tsai, “Data hiding in stained glass images,”

Proceedings of 2005 International Symposium on Intelligent Signal Processing and Communications Systems, pp. 129-132, Hong Kong, December 2005.

[23] C. Y. Hsu and W. H. Tsai, “Creation of a new type of image - circular dotted image - for data hiding by a dot overlapping scheme,” Proceedings of 2006 Conference on Computer Vision, Graphics and Image Processing, Taoyuan, Taiwan, August 2006.

[24] T. C. Wang and W. H. Tsai, “Creation of tile-overlapping mosaic images for information hiding,” Proceedings of 2007 National Computer Symposium, pp.

119-126, Taichung, Taiwan, December 2007.

[25] Y. K. Lee and L. H. Chen, “High capacity image steganographic model,” IEE Proceedings of Vision Image Signal Process, pp. 288-294, June 2000.

[26] D. L. Currie and C. E. Irvine, “Surmounting the effects of lossy compression on steganography,” Proceedings of National Information System Security Conference, pp. 194-201, Baltimore, USA, October 1996.

[27] J. Fridrich and R. Du, “Secure steganographic methods for palette images,”

Proceedings of the Information Hiding Workshop, Vol. 1768, pp. 61–76, Dresden, Germany, 2000.

[28] L. M. Marvel, C. G. Boncelet Jr. and C. T. Retter, “Spread spectrum image steganography,” IEEE Transactions on Image Processing, Vol. 8, pp. 1075–1083, 1999.

[29] J. Fridrich and M. Goljan, “Digital image steganography Using Stochastic Modulation,” Proceedings of SPIE Electronic Imaging, pp. 21–24, Santa Clara, USA, January 2003.

[30] D. C. Wu and W. H. Tsai, “A steganographic method for images by pixel-value differencing,” Pattern Recognition Letters, Vol. 24, No. 9-10, pp. 1623-1636, 2003.

[31] J. Canny, “A computational approach to edge detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 8, No. 6, 1986.

[32] P. P. Chen and W. H. Tsai, “Copyright protection of palette images by a robust lossless visible watermarking technique,” Proceedings of 5th Workshop on Digital Archives Technologies, Taipei, Taiwan, August 2006.

在文檔中基於俄羅斯方塊的馬賽克畫：一種新的藝術畫與其在資訊隱藏上的應用 (頁 90-0)