Loading...

Steganalysis of Incomplete Image Using Random Fields

Ahmadi, Aria | 2012

414 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44023 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Ghaemmaghami, Shahrokh; Gholampour, Iman
  7. Abstract:
  8. Widespread transfer of digital files over networks provides a hidden channel to transfer secret messages. Current steganalysis schemes need to work on a complete image for doing the detection job that starts when the image is entirely transferred, so are often restricted to an offline process. This restriction is serious when existence of the hidden message carriers on the network is shorter than the time required for the detection process. In this thesis, we propose a structurally fast detection method to detect the data hidden in an image passing through network. We use two of most powerful steganalysis algorithms for steganalysis of images that proposed by 1) Fridrich and Pevny and 2) Liu et al. These steganalysis algorithms are employed in a multistage process to classify clean images from images which possibly contain hidden data. We study the behavior of powerful steganalysis algorithms during the transfer of JPEG images in the progressive mode of operation. By implementating these algorithms and using the queue theory, we model our multistage system and calculate its parameters, e.g. data rate, true positive rate, true negative rate, etc. Also, we propose a new feature vector which yields a detection rate improvement of 9.61% to the Fridrich-Pevny algorithm and 9.11% to the Liu et al. algorithm. We develop a pre-warning SVM, which is capable of classifying more than 50% of the stego images, using just 12 DCT coefficients, that achieves the accuracy ratio of 93.95% for Fridrich-Pevny algorithm and 99.2% for Liu et al. algorithm. In the end, the software which implements the proposed method is explained. In addition to early detection, the accuracy ratio of the proposed system is shown to be 3.84% and 1.5% more accurate than the Fridrich-Pevny algorithm and Liu et al. algorithm, respectively, when applied to a complete image
  9. Keywords:
  10. Image Watermarking ; Multistage Steganalysis ; Progressive Image Transfer ; Online Steganalysis ; Steganography ; Steganalysis

 Digital Object List

 Bookmark

No TOC