Abstract
A novel health care X-ray image watermarking approach based on human visual model and neural network technique is proposed. The human visual model is used to generate a suitable strength of the embedded watermark and can be described in terms of four properties: entropy, frequency, luminance and texture sensitivity. The neural network technique is employed to obtain the local characteristics of the original image. In experiments for different health care X-ray images, the proposed scheme can achieve a better stego-image quality as well as extracted secret image visual quality. The proposed system can be used on the Internet to reduce difficulties that medical personnel might encounter.