直方圖位移法之滑動視窗藏密分析研究

A Steganalysis Research Using Sliding Window and Histogram-Shifting

賴芃吟(Peng-Yin Lai)

指導教授 ： 陳同孝

共同指導教授 ： 陳民枝(Jeanne Chen)

國立臺中科技大學/資訊與流通學院/資訊工程系碩士班/碩士(2014年)

https://doi.org/10.6826/NUTC.2014.00023

摘要

近年來網際網路高速發展，資料傳遞相當便利，有心人士卻利用資料隱藏技術，將秘密資料藏匿在影像中進行恐怖攻擊，造成社會動盪不安，若能透過「影像藏密分析技術」事先攔截，對人們來說是一大福音。目前多數的藏密分析技術大都透過SVM分類器擷取特徵，進行特徵分類訓練與測試，根據結果來判定「是」、「否」藏密，消耗電腦計算的時間成本不少。為有效減少藏密偵測的訓練時間，並且提高藏密分析的精準性，本研究針對2006年Ni等學者所提的直方圖位移法進行藏密分析，將影像直方圖灰階値由0到255排列，以1x4滑動視窗由左向右掃描，根據1x4滑動視窗中四個元素的特徵進行分組歸類，同時計算1x4滑動視窗中每四個元素間的相似度，找出一個最佳解當做判定影像是否有藏密的依據。本方法好處是不需要透SVM分類器，即可判斷出影像是否有藏密，取出秘密資訊之後，影像還可以正確的被還原，在主動式的藏密分析技術上，提供一個更為簡易又快速的偵測方法。本研究採用三種不同尺寸的影像圖庫進行實驗，結果證明有94.13%以上的正確率，影像一旦被偵測出有藏密，甚至有99.64%以上可以正確的被估測出藏密量及藏密位置。未來研究發展方向，希望可以針對各種不同的資料隱藏法，提供更簡易、更快速的高效能藏密分析技術，有效抑止恐怖事件的發生。

關鍵字

資料隱藏 ； 藏密分析 ； 滑動視窗藏密分析 ； 支援向量機

並列摘要

In recent years, the rapid development of the Internet has made data transfer very convenient. The use data hiding as part of terrorist movement has caused social uneasiness. However, steganalysis could be used as a way to intercept and to break possible terrorist movements. At present, most steganalysis mostly use the SVM classifier to capture feature which are then used in feature classification training and testing. The results are then used to determine a stego image or cover image. The current SVM classifier can save some computing time. In the thesis, the aim is to further reduce steganalysis detection training time and to improve the accuracy of steganalysis. The study proposed is based on Ni et al Histogram-Shifting’s which was present in 2006. The 1x4 gray value histogram of the image is scanned from left to right by a sliding window 0-255 times. The 1x4 sliding window group feature of four elements are classified, while calculating the similarity between the elements of each of the four 1x4 sliding window. An optimal solution is located from the stego image. This method does not require the benefits through SVM classifier to determine the stego image. After removing the secret information, the image can also be restored. This provide an easy and fast detection method on active Steganalysis. This study makes used of three different sizes of image gallery. Experimental results showed an accuracy detection rate of 94.13% and peak estimation rate of 99.64%. Future research includes a variety of data hiding method to provide easier and faster Steganalysis which can effectively suppress terrorist incidents.

並列關鍵字

Data hiding ； Steganalysis ； Sliding window steganalysis ； Support Vector Machine

參考文獻

[28] Lee, C. W., and Tsai, W. H., “A Lossless Data Hiding Method by Histogram Shifting Based on an Adaptive Block Division Scheme”, Pattern Recognition and Machine Vision, River Publishers, Aalborg, Denmark, pp. 1-14.

[1] Alattar, A. M., “Reversible Watermark Using the Difference Expansion of a Generalized Integer Transform,” IEEE Transactions on Image Processing, Vol. 6213, No. 8, pp. 1147-1156, 2004.

[2] Avcibas, I., Memon, N., and Sankur, B., “Steganalysis Using Image Quality 64 Metrics,” IEEE Transactions on Image Processing, Vol. 12, No. 2, pp. 221-229, 2003.

[3] Cancelli, G., Doërr, G., Cox, I., and Barni, M., “Detection of ±1 Steganography based on the Amplitude of Histogram Local Extrema,” in Proc. IEEE Int. Conf. Image Processing, San Diego, CA, 12-15. October 2008.

[4] Celik, M. U., Sharma, G., Tekalp, A. M., and Saber, E., “Lossless Generalized-LSB Data Embedding,” IEEE Transactions on Image Processing, Vol. 14, No. 2, pp. 253-266, Feb 2005.