A new passcode based approach for hiding classified information in images

This paper proposes a method of hiding classified data in images based on three different levels of security. Instead of hiding data directly, pixels are selected randomly and their higher nibble bits are matched with the data bits. A two bit code (2BC) is generated to encode the location of the matching bits. The two bit code is embedded in the image based on a password using three different suggested techniques. Simulation results on IEEE standard images of size 512 by 512 show that the suggested techniques are capable of achieving PSNRs between 56 dB and 67 dB with hiding up to 800 ASCII characters. The three different levels of security make this method highly difficult to intercept and useful for secure open channel communications.