Logo watermarking for speech signal protection in the compressive sensing scenario

The procedure for watermarking of speech signal in the compressive sensing scenario is presented in the paper. The watermark is created as a logo image, embedded into the time - frequency plane. Logo is embedded into the certain number of bit planes - each selected bit plane contains only part of the logo image in order to provide higher secrecy. The logo extraction is presented, as well. The robustness of the procedure is tested under the Compressive Sensing attack.