A New Copyright- and Privacy-Protected Image Trading System Using a Novel Steganography-Based Visual Encryption Scheme

In copyright- and privacy-protected image trading systems, the image sent to the trusted third party (TTP) is visually encrypted, and the image is traditionally unrecognizable. The image, however, is suspicious and has possibility to be attacked. Image steganography then becomes more interesting than image encryption for this application, however, applying image steganography instead of image encryption degrades the fingerprint extraction performance. In addition, the CP is allowed to directly contact with the consumer in the conventional systems. Thus, the consumer’s privacy is not protected completely. This paper then proposes a new copyright- and privacy-protected image trading system with a novel steganography-based visual encryption scheme, where the scheme protects a commercial image much more securely by generating a recognizable image instead of a suspicious encrypted image. By replacing amplitude components of a dummy image by those of a commercial image, the output image looks like a degraded dummy image instead of the commercial image, while it contains some details of the commercial image, i.e., the amplitude components of the commercial image are hidden into a dummy cover image. A discrete cosine transform-based fingerprinting method, which is compatible with the proposed amplitude component replacing scheme, is also proposed in this paper to solve the problem of applying image steganography. As another contribution of this paper, the CP is not allowed to directly send the image reconstruction key to the consumer, for further consumer’s privacy protection. The second TTP is then introduced to the proposed system, and in addition, the image reconstruction key is encrypted by the CP before being sent to the consumer via the second TTP for more security. Experimental results show that the proposed scheme generates recognizable images and perfectly visually encrypts the commercial images. It also achieves much higher reconstructed image qualities than those of the conventional scheme, and the proposed system simultaneously enhances the fingerprinting performance using the proposed compatible fingerprinting method.