Write Algorithms Extract Jpeg Images Lsb Hidden Attack

Using the basic knowledge of cryptology and statistics, and with the traditional steganalysis method, we discuss how to extract the hiding message which is embedded by LSB steganography of JPEG images.Passive attacks to steganography, mainly include the detection, extraction and restoration of hidden messages. Traditional steganalysis methods mainly focus on the detection of the existence of hidden message. The difference between detection and extraction is that: we can dectect the hiding message when the whole image is considered to be changed, while if we want to extract the hiding message, we must find out every place which has been changed. However, the study of extracting technique is so difficult that within our knowledge there seems to have few attempt in extracting steganalysis.In this paper, we discuss the extracting attack to LSB steganography of JPEG images. The main contributions of this thesis are summarized as follows:Firstly, we discuss the extracting attack to sequential LSB steganography of JPEG images. For sequentially embedding methods, the task of extraction is just to determine the beginning-point and end-point of the embedded message. We can estimate the two points by the same means, suppose that sequential LSB steganography of JPEG images embed message from the beginning of the image. In the present paper, we propose a method, based on a kind of detection steganalysis technique called Chi-square attack. Our approach can estimate the beginning-point and end-point of the embedded message accurately.Sencondly, we discuss the extracting attack to random LSB steganography of JPEG images. The point of the extracting attack to random LSB steganography of JPEG images is to recover the stego-key. In fact, extracting attack can be treated as a special kind of cryptanalysis, and we solve this problem by using both traditional steganalysis and traditional cryptanalysis. Our approach is to build measures, which have some correlation with the correct key, so the correlation attack of cryptanalysis can be used to identify the stego-key. The first is max value method, we build a set as a measure, which has correlative relationship with the key of key space. The second is non-statistical method, the process of running through the whole key space can be transformed into the comparisonof random vectors. So we can get the stego-key by using Spearman rank correlation test.