A novel steganographic system using continued fraction

This paper presents a novel steganographic system based on continued fraction for embedding text in an image. The grey levels are partitioned into disjoint sets based on continued fraction expansion. The input message is represented in binary format. Grey levels are transformed depending on the set in which it belongs and the message bits. The number of bits which can be embedded depends on the cardinality of the set. The decoding algorithm is also proposed. For comparative study variation in peak signal to noise ratio (PSNR) values and statistical parameters like variance, skewness and kurtosis with respect to different volumes of embedded information are considered. Results shows that proposed method has better performance compared with the existing methods like Least Significant Bit (LSB) manipulation and Discrete Cosine Transform (DCT) based method.