Loading...

Digital Image Forensics

Azarian-Pour, Sepideh | 2014

583 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 46301 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Massoud, Babaie Zade
  7. Abstract:
  8. In the past few decades there has been rapid advance in the use of digital cameras in different fields of art and science. Photo editing softwares have provided extensive facilities for their users and graphic softwares have astonished people with artificial yet fabulous images. Under these circumstances, recognition and distinction of authentic images from digitally-manipulated ones have become a critically important but notoriously daunting task. Users of the internet and computers need to recognize authentic images from the manipulated ones, or distinguish a composite photo from an original one. Digital image forensicsa was born as a response to these demands and has so far provided promising venues in tacking the issues pertaining to examining the authenticity of images through a number of proposed active and passive methods. The present thesis centers on passive or blind methods due to their wide range of usage and application.
    Passive methods have been categorized in six families: format-based, statistic-based, pixel-based, camera-based, geometric, and physics-based methods. JPEG ghost detection method, which is a subset of format-based approaches, estimates the quality factor of every region by scrutinizing the compression artifacts (such as double-quantization effect). The inconsistency of these quality factors indicates that the photo is a composite one created from two different cameras; and
    therefore, the photo is not valid. However, this method suffers from a number of flaws. One of the fundamental shortcomings of this method is its lack of automaticity. Another problem is that in a composite forged photo, even if the quality of the two different regions is different,
    if DCT blockings are not aligned, JPEG ghost does not appear. In this thesis we proposed a new automatic segmentation algorithm where, in the first place, the ghost borders are carefully extracted; then, they are classified in original or tampered groups by thresholding Bhattacharyya
    distance in feature space. In our proposed algorithm the second above-mentioned problem is sorted out as well, since all the 64 DCT blocking shifts are automatically analyzed. Our algorithm provides an excellent alternative to the traditional way of dealing with this problem where, in the absence of an automatic and efficient way, searching among 6400 different pictures imposed on the forensic analyzer which is inevitably subject to diverse recognition errors
  9. Keywords:
  10. Image Authentication ; Blind Image Forensics ; Double Joint Photographic Expert Group (JPEG)Compressiom ; Forgery Detection ; Malicious Manipulation ; Image Tampering

 Digital Object List

 Bookmark

No TOC