Digital image authentication based on second-order statistics

The usage of digital images has been increased nowadays and the integrity of an image become an important aspect which has to be protected from any attacks. Protection given to an image should not change the content of the image and also sensitive to modification caused by a third party (an intentional attack) yet robust to an incidental attack like noise. Multimedia Signature Scheme (MSS) is one of several schemes used to protect the integrity of an image. This scheme uses image feature to generate a signature for the image. The authentication process is done by comparing the extracted feature from the signature and the feature from received image. This paper uses a combination of MSS and a feature extraction technique called second-order feature statistic to build an authentication scheme which produces a high quality image and protects the image integrity. The signature of the image will be embedded into the image using watermarking technique. Watermarked images produced by this scheme have a good quality compared to the original image. The experimental results show the efficacy of the proposed scheme in which detection accuracy approaches 74.2%.