A Multibit Audio Steganography based on DWT and Supervised Neural Network
Steganography is a most reliable technique for information security but it has three challenges imperceptibility, capacity and robustness. Standard LSB technique is easy to implement and meets imperceptibility issue properly but suffers from less capacity and low robustness. Multi bit steganography solves this problem by incrementing capacity but it also cannot meet the robustness issues. With the increase of LSB layer the robustness increases but at the same time imperceptibility reduces. But Multi-layer steganography technique with DWT increases robustness as well as increases imperceptibility.
Here, the technique first focuses on mapping by using DWT up to 2 levels and the obtained result then undergoes with multibit embedding by considering the deviation of the target binary and the secret audio binary. The lower deviation area is taken into consideration and target binary bits are then replaced by secret binary bits.
Keywords - DWT, Cover Audio, Secret Audio, Stego Audio, Robustness, Imperceptibility, Distortion, Security.