Abstract-Steganography is a hiding system that conceals the information in a way only the sender and the recipient know about its existence. Various steganography methods developed to cover different objectives of steganography applications. All these objectives support the main goal of steganography that is undetectability. In this paper a DCT based steganography system is proposed to embed information in 4 th bits of DCT coefficients and optimize quality of the obtained stego-images applying genetic algorithm that looks for the best position for embedding. The main idea is derived from SSB-4 method which embeds the message in more significant bits to be more resistant against various steganalysis methods. The experimental results show that the proposed method enhances the imperceptibility and undetectability of the stego-images and is resistant against some steganalysis techniques such as chi-square attack.Index Terms-Discrete cosine transform, genetic algorithm, steganography, stego-image.

Feature based extraction techniques attract a great attention in digital image watermarking to survive against various removal and geometric attacks. Due to information hiding according to the image content, these techniques find a huge popularity amongst different image watermarking techniques during the recent years. In this paper a survey is investigated in feature based extraction techniques to check, evaluate and compare the recent developments in this area. The aim is to find the vital influential factors on robustness, imperceptibility, capacity, security and speed of the watermarking algorithms proposed based on feature based extraction techniques as a recommendation for future researches. The experimental results prove the robustness of the selected features in each technique against the mentioned geometric and removal attacks.

Digital image watermarking is a branch of information hiding in which the ownership information will be hidden in the cover image. Performing digital image watermarking, some important issues such as imperceptibility, robustness, capacity and security should be considered. The strength and weakness of each of these issues depends on the watermark algorithm and also the domain in which the watermark has been hidden. In this research we investigate the role of the different domains used in digital image watermarking and categorize them based on the power of the mentioned issues that each domain potentially and innately offers.