An Audio Watermarking Algorithm Using Group Quantization of DCT Coefficients

摘要

In this paper, we propose a watermark algorithm based on DCT and audio feature quantization. The proposed algorithm can be easily implemented and has a variety of variable factors to protect user information, such as quantization groups, filtering, and an audio feature extraction formula. If these factors are unknown, it is more difficult for them to be attacked. Our algorithm can extract a watermark without help from the original audio signal, giving it a higher degree of usability. The experimental results show that the proposed scheme is inaudible and robust against common signal processing techniques, including low pass filtering, noise addition, and compression. The proposed algorithm is particularly robust with respect to compression.