A robust digital watermarking scheme based on optimal coefficients selector about subimages

This paper proposes a watermarking technique based on image features. Four subimages are obtained from the host image by using subsampling. Wavelet transform is applied to the subimages. A key will be formed by comparing the DWT mid-frequency coefficients of the subimages. In contrast to conventional approaches, our method can strengthen the fidelity of images through considering local image characteristics, without sacrificing the robustness of images. Watermark extracting is simply by the key and does not need original image. Experimental results are promising. This technique is significant to medical images, military images and so on.