Adaptive Audio Steganography Based on Improved Syndrome-Trellis Codes

Syndrome-Trellis Code (STC) is a near-optimal convolutional method for adaptive steganography. Hitherto, the existing steganography commonly depends on carefully designed distortion cost function, which controls embedding position of message in cover signal. From another point view, we implement by improving STC coding process (named Adaptive-STC). The parity-check matrix key to encoding and extraction STC. In this work, prove that average change probability corresponding elements can be changed adjusting its submatrix. Following that, specially design an replace restrict position. generation formulated as multi-constrained integer programming problem width submatrix allocated at fixed height. To solve particular problem, propose targeted intelligent optimization algorithm GOAS) adaptively generate according different audio cover. experimental results show proposed outperforms state-of-the-art with reduced changes improved quality while ensuring ability against steganalysis.