資訊隱藏的研究－使用完全平方數、混合邊緣偵測與最小失真以及模數函數

Data Hiding Using the Perfect Square Number, Hybrid Edge Detector with Minimal Distortion and the Modulus Function

冷輝世(Hui-Shih Leng)

指導教授 ： 曾顯文

朝陽科技大學/資訊學院/資訊管理系/博士(2014年)

摘要

資訊隱藏是一門將資訊藏入載體並秘密傳送的技術。由於數位影像被廣泛的使用在網際網路上，且具有藏量大的特性，因此常被用來做為載體。當將機密訊息嵌入原始影像後產生偽裝影像，且此偽裝影像相較於原始影像會有失真，如何減少失真則是資訊隱藏的重要課題。資訊隱藏的評估有兩項重要的指標，藏入量與隱蔽性。藏入量是指由原始影像中平均每個像素所嵌入的機密訊息數量所決定，隱蔽性則是由計算PSNR值來評估。高隱蔽性代表原始影像與偽裝影像的差異小，亦即低失真。
資訊隱藏依其特性又分為可逆式與不可逆式兩種。可逆式是指接收方在擷取出機密訊息後可以無失真的還原原始影像。由於必須保留部份額外資訊以還原原始影像，所以藏入量比一般不可逆式低。雖然不可逆式藏量較大，但是必須符合人眼無法辨識的程度。
本論文提出三種常見的不可逆式資訊隱藏技術提出改進的方法。首先，在鄰近像素差值藏入法的量化表中使用完全平方數取代傳統的二進位值做為量化表的範圍。其次，在邊緣偵測和最低位元藏入法中使用新的邊緣檢測方式增加藏入量並以最小失真法提高偽裝影像的品質。最後，對模數函數藏入法引入新的概念以及延伸應用。

關鍵字

最低位元藏入法 ； 邊緣偵測 ； 鄰近相素差值藏入法 ； 不可逆式資訊隱藏 ； 模數函數藏入法

並列摘要

Data hiding is a technique that conceals data into a carrier for conveying the secret message confidentially. Digital images are widely transmitted over the internet and with large payload, so digital images often serve as a carrier. After embedding the secret message into the cover image, the cover image termed as stego image and distortion occur. Reduce the distortion is an important issue in data hiding. The measurement of data hiding has two requirements, payload and imperceptibility. The payload is determined by the number of secret message embedded in each pixel on the cover image. The imperceptibility is calculated by peak signal-to-noise ratio. High imperceptibility implies low distortion difference between cover image and stego image.
Data hiding has two types: reversible and irreversible. The reversible data hiding can restore the cover image without any distortion after the secret message has been extracted. Due to requirement of extra information of reversible data hiding, it has lower payload than most irreversible data hiding schemes. Despite the irreversible data hiding has a higher payload; it must comply with the human visual system which implies the stego image cannot be recognized by human eyes.
In this dissertation, we propose three methods of irreversible data hiding. First, we use the perfect square number divide the quantization range table on the pixel-value-differencing scheme. Second, we use a new hybrid edge detector increase the embedding capacity by least-significant-bit substitution scheme and use the minimal distortion method to improve the quality of the stego image. Finally, we try to use the geometry relation on the modulus function to replace the calculation complexity and extend the exploiting-modification-direction scheme and the fully-exploiting-modification-direction scheme to n-dimensional hypercube.

並列關鍵字

hybrid edge detector ； pixel-value-differencing ； Irreversible data hiding ； least-significant-bit substitution ； modulus function

參考文獻

[41] H. W. Tseng & H. S. Leng, "A Stegnographic method based on pixel-value differencing and the perfect square number," Journal of Applied Mathematics, 2013, Vol. 2013, Article ID 189706.

[1] D. C. Wu & W. H. Tsai, "A steganographic method for images by pixel-value differencing," Pattern Recognition Letters, 2003, Vol. 24, No. 9-10, pp. 1613-1626.

[2] H. C. Wu, N. I. Wu, C. S. Tsai & M. S. Huang, "Image steganographic scheme based on pixel-value differencing and LSB replacement methods," IEEE Proceedings Vision Image and Signal Processing, 2005, Vol. 152, No. 5, pp. 611-615.

[3] C. H. Yang & C. Y. Weng, "A steganographic method for digital images by multi-pixel differencing," Proceedings of International Computer Symposium, 2006, pp. 831-836.

[4] K. H. Jung, K. J. Ha & K. Y. Yoo, "Image data hiding method based on multi-pixel differencing and LSB substitution methods," International conference on Convergence and Hybrid Information Technology, 2006, pp. 355-358.