Adaptive chosen-ciphertext attack on secure arithmetic coding
IEEE Transactions on Signal Processing
| Hi-index | 0.00 |
A semi-adaptive time variant modeling method is developed to be used in conjunction with an arithmetic coder to compress data in a secure fashion. The modeling method introduces randomness into the symbols coding probabilities by ensuring that the update time for computing the coding probabilities is performed at random intervals as specified by the output of a stream cipher. The method results in the ability to effectively compress the data while ensuring its security. The compression ratios of the method are very close to the ratios obtained from standard implementation of adaptive modelers. The computational complexity of the proposed method is less than the computational complexity of performing a compression step that is followed by an encryption step. Additionally, the computational complexity of the method is less than the computational complexity of typical implementations of adaptive modelers. The methodology allows for building high security architectures that are ideal for real time operations.