DCTおよびVQを用いた画像電子透かし

Abstract

This paper describes two digital watermarking methods using DCT (Discrete Cosine Transform) and VQ (Vector Quantization). A bi-level (1 bit/pixel) watermarking image has been embedded in target gray scale images. In the DCT watermarking method, DCT coefficients of target gray scale images are modified by watermarking information 1/0. In the VQ watermarking method, a VQ code book is divided to 2 groups which are assigned to watermarking information 1/0, respectively. Experiments have been carried out. The SNR (Signal to Noise Ratio) of embedded target images and the BER (Bit Error Rate) of extracted watermarking information have been studied. Our Experimental results have shown that the DCT method is superior to the VQ method.