Image authentication based on chaotic system with feedback and palm characteristics

An image authentication scheme is proposed based on a chaotic system with feedback and palm characteristics. With a Trusted Third Party (TTP), the integrity of an image can be detected, meanwhile, the sender of the image can be identified reliably. During embedding period, an image is first divided into non-overlapped blocks. An authentication code is produced based on two adjacent blocks using a chaotic system with feedback. To identify the sender of an image and prevent denying event, biometrics information of the sender is applied in our scheme. A feature vector of palm is extracted and converted to a stream of bits. Whereafter, the authentication codes and the palm feature bits are encrypted using the public key of TTP. Then the encrypted information are transmitted to TTP, who will decrypt the received message and interleave them in a secret manner. The interleaved message is given back to the sender, and embedded in the corresponding image blocks. During authentication, a user detects the image integrity and localizes the tampered regions. At TTP, the sender is identified because of the reliable palm recognition. If the sender denies an image, TTP can capture his palm image on site, extract a feature vector, and compare it with the feature hidden in the image to confirm identity. Experiments reveal the effectiveness of our scheme.