Steganography - SEC302

Neil F. Johnson. Steganography. Technical Report. November 1995.

Kurak and McHugh identify two kinds of compression, lossless and lossy [Kurak92]. Both methods save storage space but may present different results when the information is uncompressed.

Lossless compression is preferred when there is a requirement that the original information remain intact (as with steganographic images). The original message can be reconstructed exactly. This type of compression is typical in GIF2 and BMP3 images.
Lossy compression, while also saving space, may not maintain the integrity of the original image. This method is typical in JPG4 images and yields very good compression.

To illustrate the advantage of lossy compression, Renoir's Le Moulin de la Galette was retrieved as a 175,808 byte JPG image 1073 x 790 pixels with 16 million possible colors. The colors were maintained when converting it to a 24-bit BMP file but the file size became 2,649,019 bytes! Converting again to a GIF file, the colors were reduced to 256 colors (8-bit) and the new file is 775,252 bytes. The 256 color image is a very good approximation of Renoir's painting.

____________________
    2     Graphic Interchange Format developed by Compuserve to be a
          device-independent method of storing images.
    3     Windows and OS/2 bitmap picture file.
    4     Joint Photography experts Group (JPG/JPEG) is a device-
          independent method for storing images which supports 24-bit
          images.

(go back)

Next Section: 3.2 Evaluation Method