A wavelet domain LSB insertion algorithm for high capacity audio steganography

摘要

In this work, we increased the capacity of the classic LSB insertion method by performing the embedding process in the wavelet domain. The algorithm uses perfect reconstruction filterbanks and embeds additional information inside wavelet domain of audio signal by modifying LSB values of wavelet coefficients. Objective and subjective tests show large advantage of the proposed method over time insertion method. For the same SNR values, proposed algorithm outperforms classical LSB algorithm by 150-200 kbps of hidden data. Subjective experiments showed that wavelet domain information hiding scheme is acoustically more transparent as well.