The chances of data getting lost while it is being transferred between the sender and receiver is very high these days. Since these data are very sensitive whose security cannot be compromised, there is a need for highly secure systems to transfer the data without compromising the content and the quality. Steganography techniques help us to achieve these objectives. In the present time, various organizations and industries are using cover image arbitrarily. In such cases, it does not provide personalized approach to the whole process of data hiding. Thus, as a result of this limitation, there is motivation to build such an application in which the system selects which image is most suitable for hiding data accordingly. The main algorithm being used here is the regression algorithm which consists of other algorithms like linear regression, decision tree regression. This application also extracts the hidden data from the generated stego image.
TopLiterature Survey
Khandare et al. (1996) discuss about data hiding technique called Steganography. Their paper presents a survey on various data hiding techniques in steganography along with the comparative analysis of these techniques. Their paper presents both traditional and novel techniques for addressing the data-hiding process and evaluates these techniques in light of three applications: copyright protection, tamper proofing, and augmentation data embedding. The authors have discussed about steganography and presented some notable differences between steganography and cryptography. They also surveyed various data hiding techniques in steganography and provided a comparative analysis of these techniques.
Goel et al. (2013) attempted to work on most of the prominent algorithms. Their work was mostly related to finding out the parameters that steganography can be based on and their comparison. Their paper deals with hiding information using LSB, DCT and DWT techniques. These algorithms are measured based on their MSE and PSNR values. Along with these, two other parameters are used to measure the effectiveness of these algorithms. They are Robustness and Capacity payload. The authors have successfully implemented the algorithms proposed in this paper and the results have been documented. From their results, they came to the conclusion that DCT having higher PSNR provides best quality of images.
Chen Ming et al. (2006) made a comparative study on different steganalysis techniques. They have implemented Markov Chains to try and test various techniques and made their conclusions.
Zöllner et al. (1998) studied SVM and its variation RBF to achieve the best possible results. According to his findings, the best result obtained was the usage of RBF as it was automatic and very versatile.
Joachims et al. (2000) has studied the effect steganography has on certain image features after the data has been embedded. They found out that certain statistical measures are different for cover images and stego images.