Abstract:
This research proposes to use a raw speech as watermark signal in the digitalwatermarking process, together with the detection of recovered watermark signal usinghuman hearing system. The wavelet packet transform based technique is used to embedthe raw speech signal into a video sequence. The experimental results show that usinglistining base technique to detect the watermark is more robust against various types ofattacks compared to the traditional detection technique by characteristics base that isbrightness/contrast enhancement/decrement, Gaussian distributed noise adding, JPEGstandard compression, high-pass filtering, MPEG-2 encoding, image resizing and framecutting.