پنهان‌شكني در تصاوير با استفاده از ماتريس هم‌رخدادي و شبكه عصبي
محورهای موضوعی : مهندسی برق و کامپیوترصدیقه قنبری 1 * , نجمه قنبری 2 , منیژه کشتگری 3 , سیدحسن نبوی کریزی 4
1 - دانشگاه شيراز
2 - دانشگاه آزاد اسلامی واحد زاهدان
3 - دانشگاه صنعتي شيراز
4 - آموزشكده فني حرفه‌اي شهيد منتظري مشهد
کلید واژه: پنهان‌شكني پنهان‌نگاري شبکه عصبی ماتریس هم‌رخدادی,
چکیده مقاله :
پنهان‌نگاری یا استگانوگرافی هنر برقراری ارتباط پنهانی است و هدف آن پنهان‌کردن ارتباط به‌وسیله قراردادن پیام در یک رسانه پوششی می‌باشد و پنهان‌شکنی هنر کشف حضور اطلاعات پنهان است. ماتريس هم‌رخدادي تصاوير (GLCM) ماتريسي است كه در بردارنده اطلاعاتي در رابطه با ارتباط بين مقادير پيكسل‌هاي مجاور در يك تصوير مي‌باشد. در اين تحقیق به كمك بررسي و تحليل ماتريس هم‌رخدادي در تصاوير پوشانه (تصويري كه حامل اطلاعات نباشد) و گنجانه (تصويري كه حامل اطلاعات باشد)، الگوریتمی ارائه می‌گردد تا بتوان تصاویر گنجانه را تشخیص داد. در الگوریتم پیشنهادی ابتدا با استفاده از روش LSB اطلاعات در تصویر مورد نظر پنهان می‌شود، سپس به استخراج ویژگی‌هایی از ماتريس GLCM تصویر پوشانه و گنجانه می‌پردازیم که در این دو تصویر متفاوت باشند. ویژگی‌های استخراج‌شده برای آموزش شبکه عصبی مورد استفاده قرار می‌گیرد. این الگوریتم بر روی تصاویر پایگاه داده‌هاي استاندارد تست گردیده و موفقیت آن 83% می‌باشد.
Steganography is the art of hidden writing and secret communication. The goal of steganography is to hide the presence of information in other information. steganalysis is the art and science of detecting messages hidden using steganography. Co-occurrence matrix is the matrix containing information about the relationship between values of adjacent pixel in an image. In this paper, we extract features from Gray Level C0-occurrense Matrix (GLCM) that are difference between cover image (image without hidden information) and stego image (image with hidden information). In the proposed algorithm, first, we use a combined method of steganography based on both location and conversion to hide the information in the image. Then, using GLCM matrix properties, we investigate some difference values in the GLCM of the cover and stego images. We can extract features that were different between cover and stego images. Features are used for training neural network. This algorithm was tested on 800 standard image databases and it can detect 83% of stego images.
[1] C. S. Shieh, H. C. Huang, F. H. Wang, and J. S. Pan, "Genetic watermarking based on transform - domain techniques," Pattern Recognition, vol. 37, no. 3, pp. 555-565, Mar. 2004.
[2] M. Mahdavi, S. Samavi, N. Zaker, and M. Modarres - Hashemi, "Steganalysis method for LSB replacement based on local gradient of image histogram," Iranian J. of Electrical & Electronic Engineering, vol. 4, no. 3, pp. 59-70, Jul. 2008.
[3] A. Westfeld and A. Pfitzman, "Attacks on steganographic systems," in Proc. of 3rd Int. Information Hiding Workshop, Springer-Verlag, pp. 61-76, Sep. 1999.
[4] J. Fridrich, M. Goljan, and D. Holga, "Steganalysis of JPEG images: breaking the F5 algorithm," in Lecture Notes in Computer Science, Springer, vol. 2578, pp. 310-323, Oct. 2002.
[5] J. Fridrich, R. Du, and L. Meng, "Steganalysis of LSB encoding in color images," in Proc. of IEEE International Conf. on Multimedia, vol. 3, pp. 1279-1282, Jul. 2000.
[6] J. Fridrich, M. Goljan, and R. Du, "Reliable detection of LSB steganography in grayscale and color images," in Proc. of ACM, Special Session on Multimedia Security and Watermarking, pp. 27-30, Otawwa, Canada, Oct. 2001.
[7] S. Dumitrescu, X. Wu, and N. Memon, "On steganalysis of random LSB embedding in continuous-tone images," in Proc. IEEE Int. Conf. on Image Processing, vol. 3, pp. 641-644, 24-28 June 2002.
[8] A. Yadollahpour and H. Miar Naeimi, "Attack on LSB steganography in color and grayscale images using autocorrelation coefficients," European J. of Scientific Research, vol. 31, no. 2, pp. 172-183, May 2009.
[9] T. Zhang, Y. Zhanh, X. Ping, and M. Song, "Detection of LSB steganography based on image smoothness," in Proc. IEEE Int. Conf. Multimedia and Expo, vol. 1, pp. 1377-1380, Jul. 2006.
[10] N. Provos, "Defending against statistical steganalysis proc," in Proc. 10th Usenix Security Symp, Usenix Assoc., vol. 10, pp. 323-335, Aug. 2001.
[11] J. Fridrich, M. Goljan, and D. Hogea, "Attacking the outguess," in Proc. ACM Workshop Multimedia and Security 2002, Dec. 2002.
[12]
[13] J. Fridrich, "Feature - based steganalysis for JPEG images and its implications for future design of steganographic schemes," in Proc. 6th Information Hiding Workshop, LNCS, Springer-Verlag, vol. 3200, pp. 67-81, May 2004.
[14] J. Fridrich, M. Goljan, and D. Hogea, "New methodology for breaking steganographic techniques for JPEGs," in Proc. SPIE Electronic Imaging, vol. 5020, pp. 143-155, Santa Clara, CA, US, Jan. 2003.
[15] V. Sabeti, S. Samavi, and M. Mahdavi, "Steganalysis of embedding in difference of image pixel pairsby neural network," ISC Int'l J. of Information Security, vol. 1, no. 1, pp. 17-26, Jan. 2009.
[16] L. Shaohui, Y. Hongxun, and G. Wen, "Neural network based steganalysis in still images," in Proc. Int. Conf. on Multimedia and Expo, ICME 2003, vol. 2, pp. 509-512, Jul. 2003.
[17] P. Lafferty and F. Ahmed, "Texture based steganalysis: results for color images," in Proc. SPIE, Mathematics of Data/Image Coding, vol. 5561, pp. 145-151, Aug. 2004.
[18] Z. Sun, M. Hui, and C. Guan, "Steganalysis based on co - occurrence matrix of differential image," in Proc. Int. Conf. on Intelligent Information Hiding and Multimedia Signal Processing, pp. 1097-1100, Aug. 2008.
[19] H. B. Kekre, A. A. Athawale, and S. A. Patki, "Steganalysis of LSB embedded images using gray level co - occurrence matrix," Int. J. of Image Processing, IJIP, vol. 5, no. 1, pp. 711-720, 2011.
[20] Standard Database of USC - SIPI (University of Southern California Signal & Image Processing Institute), available at http://sipi.usc.edu/database/index.php
[21] Standard Database of BSD (BerkeleySegmentationDataset) available at http://www.eecs.berkeley.edu/Research/Projects/CS/vision/ grouping/fg/