DCT domain Stegasvm-shifted LSB model for highly imperceptible and robust cover-image / Hanizan Shaker Hussain

The importance of information security in protecting data and information has increased due to the increased use of computers and the Internet. Similarly, with one of its exciting subfields i.e. information hiding. Information hiding is a technology where the secret-messages are hidden inside other files (e.g image files). One of the areas that are popular now applying this technology is digital image steganography (image steganography). In image steganography, the most popular and widely usedtechniques is the least significant bit (LSB) that hide data into a cover-image in a spatial and discrete cosine transform (DCT) domain as well. Beside the LSB technique, there is other technique that is also influential i.e support vector machine (SVM) normally used to strengthen the embedding algorithm. Whatever techniques used in the image steganography field,the main purpose is to keep the existence of the secret-message secret. But many of the techniques previously proposed have failed to attain this main purpose. The primary concern that contribute to this problem is the non-random changes on a cover-image that constantly occurred after the embedding process. Secondly, the non-robustness of embedding algorithm to image processing operation. Therefore in this research, the new model is proposed called StegaSVMShifted LSB model in DCT domain to preserve the imperceptibility and increase the robustness of stego-images. The StegaSVM-Shifted LSB model that has been proposed that utilize HVS and embedding technique through Shifted LSB showed a good performance. This can be seen when PSNR record high value, where it displays a good quality cover-image with 48.94dB while high quality robustness for secretmessage with NC value is about 1.0. Therefore, StegaSVM-Shifted LSB model were acceptable in which it shows a higher quality steganography, thus enhancing the performance of existing works. In extracting process, by exploiting the SVM learning ability, the right secret-bits can be recovered.