Tamper Detection in RFID Tags based on Watermarking using Neural Network Hash Function

Nowadays, as tampering attacks getting more attention, data protection Radio Frequency Identification tag (RFID) becomes more important. The watermarking approach prevents unauthorized changes on the content of such labels. Due to the limitations of such tags like their simple structure, scarcity of memory bits and its binary content, using conventional methods likes watermarking and hash function are impossible. Thus, the approaches that can be applied to binary data with low size are used. This paper is based on a special algorithm of neural networks that is used to create the hash and the watermarking code that causes fewer number of valuable bits of memory to be consumed. In the proposed algorithm, all watermarking bits are securely protected and locations of the watermark bits is not detectable. This method not only has no need for secrecy watermarking algorithms, but also has other benefits such as ease and speed of implementation. Another advantage of the proposed approach is the two-step algorithm with its keys and using pseudo-random bits location makes watermark's probability of detection very difficult and inaccessible, and also increase the robustness of the method.