Intelligent detection of LSB stego anomalies in images using soft computing paradigms

Techniques for information hiding and steganography are becoming increasingly more sophisticated and widespread. With high-resolution digital images as carriers, detecting hidden messages is also becoming considerably more difficult. In this paper, we describe a universal approach to steganalyse the least significant bit steganography method for detecting the presence of hidden messages embedded within digital images. The proposed work uses the 27 features that are calculated from the three different statistical moments i.e., PDF, CF and Absolute moment calculated from wavelet multi-resolution representation of the images. We have presented the efficacy of our approach on a large collection of images, and on four different LSB steganographic embedding algorithms. Four soft computing techniques viz., Support Vector Machine, Nai¿ve Bayes classifier, Decision Tree Classifier and K-nearest neighbor classifier are employed to efficiently distinguish the pure image from the anomalous or stego image file. The proposed steganalysis algorithm can steadily achieve a correct classification rate of over 90% thus indicating significant achievement in steganalysis.