Research Of Spatial Domain Steganographic Algorithm Based On Camera Features

Information hiding, which is an old art and science of transmitting secret data by hiding it in common used and barely noticed covers, has been more and more widely used in secure communication, data integrity verification, digital copyright protection, etc in many aspects of daily life with the rapid development of computer technology and internet. Compared with the traditional cryptography method of protecting secret data, information hiding technology conceals the messages as well as presentence of the hiding data process exclusively. Stegnaography, one of the most important parts of information hiding, has attracted lots of attention over the past few years in the field of information security.The paper primarily discusses spatial domain steganography under the cover of digital images. On analyzing current classic spatial domain steganography, two spatial domain steganographic algorithm are put forward with respect to the camera feature resulted from image capturing process.Lens distortion is evitable while taking pictures with a camera, whose instinct parameters and distortion parameters can be estimated by traditional calibration. Thus camera generalized images can be undistorted to some extent unperceived. Since the lens distortion in the vicinity of the optical center of the lens is less obvious than that of being farther from optical center, therefore, in the undistort process, the amount of a pixel change whose position is far away from optical center is larger compared with the pixel near the optical center. During the undistorting process, different location in the image resulting different change amount can be used to embed different amount of secret messages in each pixel adaptively.In most consumer digital cameras, the true color image is interpolated from the image captured by a color filter array (CFA). There exists some difference between the restoration of missing data and the original complete data, which can be taken advantage of embedding information in spatial domain. Therefore, a new spatial domain steganographic algorithm is proposed based on CFA interpolation. The algorithm first estimates the CFA interpolation coefficient and CFA interpolation model. Selecting the CFA interpolation mode by computing peak signal to noise ratio, we can re-interpolate the image meanwhile adaptively embed different amount of secret messages in each pixel. The conclusion of camera features based steganographic algorithm and prospect of steganography is given at the end of this paper.