• /documents:
  • contains documents concerning jxrlib (almost exclusively own notes as there exists no proper documentation) and the ITU-T reference implementation T.835.
• /double_compression_detection
  • contains the implementation of the actual double compression detection using the ITU-T JPEG-XR reference implementation (subfolder T.835)
• /double_compression_detection/test_images
  • contains the TIF test images which are converted to JXR images
• /jpegxr_libraries:
  • contains snapshots of both libraries without modifications as zip files.
• /papers
  • contains all papers we read concerning existing double compression research (e.g. for JPEG)

• All modifications to the T.835 reference implementation are documented by a comment starting with 'ADDED MS_PS'

• Unixoide operating system (currently only tested on Linux machines)
• bison, flex, gnuplot and gnuplot-x11 have to be installed

• Open a UNIX shell
• Navigate into the root directory of this project
• Issue the following commands in the root directory of this project (in order or just the one you need)
  • make - build the T.835 project and the double compression detection project
  • make images - generate all the JXR test images (located in the subdirectory double_compression/test_images/extracted/compressed/)
  • make coefficient_data - extract the coefficient data for each JXR test image while decoding it
  • make histograms - generate/plot the histograms for each coefficient for each JXR test image
  • make quantifications - quantify the coefficient data for each JXR test image
  • make clustering - cluster the quantification data for each JXR test image by coefficient, quantification method and first compression factor
  • make clean - clean the whole project including all the JXR test images

The savest way to fully build and run the project using the test images is to issue the following command(s) in this order in the root directory of the project:

1. cd double_compression_detection
2. make clean
3. make images
4. make coefficient_data
5. make quantifications
6. make clustering