Data Hiding Techniques In Digital Images

With literally millions of images moving on the Internet each year it is safe to say that hiding data in digital image is of real concern to many in the security field. Therefore how to protect secret messages during transmission becomes an important issue. Hiding data provides a good layer of protection on the secret message, so the purpose of this thesis is to study the data hiding techniques in digital images as a new and powerful technology capable of solving important practical problems. Depending on what information in which form is hidden in the digital images, one can distinguish two types of data hiding techniques, spatial domain techniques, and frequency domain techniques. In the spatial domain techniques, a digital image serves as a carrier for a secret message. For instance, by replacing the least significant bit of each pixel in the carrier image with the secret message after changing it to stream of bits, the changes to the carrier image will be imperceptible and the secret message will be masked by carrier image. In this side, two programs had been implemented using MATLAB program to illustrate the main idea involved in least significant technique (low bit encoding), and the other to illustrate the masking technique inside the carrier image. In the frequency domain techniques, a short message is embedded in the carrier image in a robust algorithm. Robustness means the ability to survive common image processing operations, such as lossy compression, filtering, noise adding, geometrical transformations, etc. So in this technique two programs had been implemented to illustrate the main idea involved in frequency domain, one with the Fast Fourier Transform (FFT), and the other with the Discrete Wavelet Transform (DWT). After all these studies, one of the algorithms in the masking technique is developed and implemented using JAVA program to embed message into true color image with a good quality and higher capacity. Beside that the carrier images in different techniques were examined by exposing them to common signal processing operations such as image resizing, rotation, histogram equalization, lossy Compression, and Gaussian noise addition to illustrate the characteristic of the data hiding techniques, such as hiding capacity, robustness, undetectability, and perceptual transparency Finally, it has been shown that the frequency transformation techniques are more robust, and hence suitable for water marking and data hiding purpose. The spatial domain techniques exhibit loss robustness but due to its higher capacity and good quality are perfect for data hiding purposes.

Text FK_2003_56.pdf Download (807kB)

Actions (login required)

View Item