Visual cryptography and steganography in high quality digital images and audio files

Resumo

Visual cryptography is a technique that consists in dividing an image into other n images (denominated shares), so that only with a superposition of k ? n of them the input image can be reconstructed. In the method proposed by Naor and Shamir, this superposition can be digitally made, through software, or physically, printing the shares in transparencies and then stacking them. However, it is limited to operating with binary images. In this work, a technique of visual cryptography specific for the digital context is developed, with the advantage of operating in monochromatic and colored images, given that little modifications would be necessary to functioning in deeper images and even audio files. The similarities between the original and reconstructed images from the shares were superior to 97% in all the experiments conducted. In addition, techniques of steganography are employed so that the shares have visual meaning, differently from what happens with the method proposed by Naor and Shamir, in which they are visually random. Both approaches were compared and the proposed method demonstrated to be robust to digital applications.