Finding cellular automata pre-processing masks for enhancing file compression
Summary
We investigate the use of a CA generated mask to pre-process a file before compression to increase the file's compressibility. For each original-file to be compressed, the algorithm searches for an appropriate CA mask, uses the mask to pre-process the original-file into an intermediate-file, and then compresses the intermediate-file into a compressed-file which is saved with an encoding of the CA mask. To restore the original-file, the CA mask is created from the attached encoding, the compressed-file is decompressed into the intermediate-file, and the CA mask is applied to the intermediate-file to re-create the original-file. The actual compression is performed using existing compression algorithms. Our contribution is the pre-processing method which depends on the fact that the unique structure of CA allows us to encode a complex mask in a very small amount of space.Preliminary work indicates that, if an appropriate CA mask is found for a given original-file, this method can improve the compression ratio for the original-file. The amount of improvement achieved depends on finding an appropriate CA mask for each file to be compressed. The task of finding an appropriate mask and doing so efficiently is a non-trivial and challenging problem that must be solved before this method can be successful. Thus, the primary research questions of this project focus on understanding the solution space of CA masks and how to efficiently search this solution space for effective masks.
