基于心理声学模型频域能量量化的音频水印算法

Audio watermarking algorithm with quantized frequency domain energy based on psychoacoustic model

首先将心理声学模型引入水印的嵌入和提取过程,作为选择和寻求合适的频域子带进行水印嵌入和提取的指导,并在此基础上提出了一种音频频域子带能量量化的水印算法.算法利用心理学模型逐帧计算音频帧各个关键频带的全局掩蔽阈值,并以此作为参考,选择其中几个关键子带进行水印嵌入,嵌入的方式采用基于能量的量化方法.

Firstly,psychoacoustic model is introduced into watermark embedding and extraction process as the guidance of selecting and seeking the appropriate frequency domain for watermark embedding and extraction,and on this basis,a watermarking algorithm with audio frequency domain sub-band energy quantized is proposed.In this algorithm,the global masking thresholds of each key frequency domain are calculated frame by frame with psychological model and the results are taken as references,several key sub-bands are selected for watermark embedding.The embedding method adopts energy-based quantization method.