基于心理声学模型的音频隐藏图像算法

提出一种利用人耳的掩蔽效应，将图像隐藏于音频信号中，实现图像在音频中隐秘传输的算法。算法将音频信号分帧进行心理声学模型分析，估计频域中各子带的最小掩蔽阈值；然后将图像转化为比特流隐藏于音频信号频域子带，并用掩蔽阈值控制嵌入强度。仿真结果表明，隐藏有图像的音频信号主观无感知失真且图像具有不可见性，对加噪、压缩、滤波和重采样等攻击具有良好的鲁棒性。

Algorithm of hiding image in audio based on psychoacoustic model

An algorithm was presented, which hided the image in audio signal based on the masking effect, the image could be transmitted in audio secretly. Audio signal was segmented into frames and the masking thresholds of all sub-bands were estimated by psychoacoustic analysis. And the image bits were hidden in the sub-bands below the masking thresholds, which made the audio signal little perceptual distortion and the image invisible. Simulation results show that the algorithm is robust against attacks of noise, compression, filtering and resampling.