Existing Methods
Home Introduction Existing Methods SSIS Details Results and Examples Code Bibliography

 

Overview of Existing Steganographic Methods

Following is a generic block diagram of a blind stegosystem.

A message is embedded into the image by the stegosystem encoder via a secret key or password. The resulting stegoimage is transmitted over a channel to the receiver, where it is processed by the stegosystem decoder using the same key or password. During transmission, the stegoimage can be monitored by hostile factors, that must not be able to identify any presence of a secret message, but only an innocuous image.

We will briefly describe some steganographic methods which are in use today.

 

Least Significant Bit manipulation

The idea behind this scheme is to replace the LSB of the image pixels with the message bits. Most steganographic programs available today use some variation of LSB manipulation in order to hide messages in an image. The major advantage of the method is its quickness and easiness. Typically, it achieves high payload and low perceptibility. However, a too naive implementation may be susceptible to various visual and statistical attacks (for more information, see [2]-[4]). The scheme also has a very low resistance to removal. For more details, please refer to the LSB examples subsection.

 

Statistical approach

An algorithm by Bender et al. [5], called patchwork,  works by altering the statistics of the image. It pseudorandomly chooses pairs of patches (groups of pixels), then it brightens one patch and darkens the other one. This way a specific statistics, one that has a Gaussian distribution, is embedded into the image. This scheme falls into the category of watermarking techniques.

 

Spread Spectrum

Widely used in RF (Radio Frequency) communication, the technique actually spreads the bandwidth of a narrowband signal across wide band of frequencies. After spreading, the energy of the narrowband signal at any given frequency is low, and therefore difficult to detect (and jam). Several methods that utilize the principles of spread spectrum have been described in [6].

We will present and describe in detail an algorithm suggested by Marvel et al. in [1], which uses techniques of error-correcting coding, image restoration and spread spectrum communication in order to construct a blind steganographic system.