Abstract: An adaptive audio steganography of integer discrete wavelet packet domain based on auditory model is proposed. After the audio carrier is segmented, the candidate segments are decomposed by integer lifting wavelet packet. According to the sub-band masking threshold which is calculated by auditory model and the characteristics of wavelet coefficients, the embedding strength value is set adaptively. Logistic chaotic map is utilized to process the secret information. Experimental results show that besides high capability under the condition of imperceptivity, this method has high security and low error rate.

Key words: audio steganography, adaptive, auditory model, integer wavelet packet transform, chaotic sequence

摘要： 提出一种基于听觉模型的整数小波包域自适应音频密写方法。将音频载体分段后，采用整数提升小波包对候选段进行小波包分解。由听觉模型计算掩蔽阈值，根据子带掩蔽阈值和小波系数的特点对小波系数自适应地确定嵌入强度。引入Logistic混沌映射对秘密信息预处理。实验结果表明，该方法在保证不可感知的条件下嵌入容量大，安全性高，提取误差低。

关键词: 音频密写, 自适应, 听觉模型, 整数小波包变换, 混沌序列

CLC Number: 

• TP309.2