Security Authentication Method of Speech Perceptual Hashing Based on Fuzzy Commitment Scheme

Existing speech perceptual hashing authentication algorithms generally encrypt the perceptual hash value to protect the security of the algorithms. But under the principle of Kerckhoffs, the algorithm becomes transparent and fragile. And it can be seen that the secret key may be estimated when the number of times that reuses the secret key reache to a limit through the safety analysis of Shannon unicity distance. To solve these problems, we present a novel security authentication scheme of speech perceptual hashing based on the fuzzy commitment scheme (FCS). Firstly, in sender, the randomly selected BCH code word and the extracted perceptual hash value are sent to the Cmt() function to calculate and the final secure perceptual hash value (h(c), delta) can be obtained. Secondly, in receiver, the perceptual hash value is calculated again from the received speech information and then the calculated perceptual hash value with the received information. (commitment) are sent to the Decmt() function to calculate to get the h(c'). Finally, the h(c) and h(c') are matched. The experimental results show that the proposed scheme effectively avoids the probability that an attacker obtains the plaintext/ciphertext pairs without affecting the performance of original algorithm, and at the same time it ensures the security of perceptual hashing algorithm.