Robust Audio Hash Function Based on Higher-Order Cumulants

Perceptual hash function provides fast and reliable means for protection, management, and indexing of audio content. In this paper, a new audio hash function based on higher-order cumulants of coarse wavelet coefficients is proposed. The Gaussian noise suppression property and high computing complexity of higher-order cumulants are utilized to achieve the perceptual robustness, collision-free and one-wayness requirements of the proposed hash function. Experimental results demonstrate the effectiveness of the proposed hash function in terms of perceptual robustness, discrimination and collision-free.