Loading...

Image Steganalysis Based on Sparse Representation

Jalali, Arash | 2015

483 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 47755 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Ghaemmaghami, Shahrokh
  7. Abstract:
  8. This thesis explored and described a new steganalysis system modeling. Image steganalysis systems are divided into two parts: feature extraction from images and classification of images. Many researches have been done in both parts and satisfactory results have also been reported. There are acceptable steganalysis methods which work accurately in high rates of steganography; however steganography in low rates is still undetectable. By lowering the rate of steganography in images, the difference between stego images and clean images would be reduced which accordingly led to the reduction of the difference between the corresponding extracted features. Thus, the ability of classification accuracy used in steganalysis, is very important in identification of stego images with low rate of hidden data. One of the most powerful classifiers is sparse representation classifier which shows a great accuracy and differentiation in many applications. Compared to other conventional classifiers, sparse representation classifier has a high computational load. This thesis proposed a new method which applies both double sparse representation and differentiating sparse dictionary learning. In classifications based on sparse representation, besides using of powerful sparse classifiers, computational load will be decreased a lot and as a result a high accurate and practical steganalysis system will be achieved. Experimental results show a high accuracy in low rates of steganography. Also computational load in this method will be decreased five times lower than general sparse classifiers
  9. Keywords:
  10. Steganography ; Steganalysis ; Dictionary Learning ; Sparse Representation ; Double Sparse Representation

 Digital Object List

 Bookmark

No TOC