An Audio Information Hiding Algorithm Based on ICA

A novel audio information hiding technology combined with ICA (independent component analysis) and QIM (quantized index modulation) is proposed. ICA method is applied to audio signal processing to obtain the statistical independent sources as hiding channels. As ICA processing is sensitive to the only dividing matrix, security is guaranteed. A piece of meaningful speech signal is hidden in the carrier audio. Through analytic comparisons and experimental results, this algorithm is found to have a larger hiding capacity and a lower distortion to carrier audio, and also robust against a variety of common signal processing manipulations