Hungarian-Puzzled Text with Dynamic Quadratic Embedding Steganography

Abstract

Least-Significant-Bit (LSB) is one of the popular and frequently used steganography techniques to hide a secret message in a digital medium. Its popularity is due to its simplicity in implementation and ease of use. However, such simplicity comes with vulnerabilities. An embedded secret message using the traditional LSB insertion is easily decodable when the stego image is suspected to be hiding a secret message.  In this paper, we propose a novel secure and high quality LSB embedding technique. The security of the embedded payload is employed through introducing a novel quadratic embedding sequence. The embedding technique is also text dependent and has non-bounded inputs, making the possibilities of decoding infinite. Due to the exponential growth of and quadratic embedding, a novel cyclic technique is also introduced for the sequence that goes beyond the limits of the cover medium. The proposed method also aims to reduce the noise arising from embedding the secret message by reducing bits changed. This is done by partitioning the cover medium and the secret message into N partitions and artificially creating an assignment problem based on bit change criteria. The assignment problem will be solved using the Hungarian algorithm that will puzzle the secret message partition for an overall least bit change.