A New Approach to File Encryption and Compression

In this paper, a new compression/encryption scheme is presented. The basic idea of this scheme is to represent a monotonically increasing sequence of integers with a single integer such that the original sequence can be restored perfectly. First, we recognize the binary bit patterns from the input file using Lempel-Ziv incremental parsing algorithm. Then a second pass through the file revises the pattern repertoire, replacing contiguous smaller patterns with equivalent bigger pattern wherever possible. Then we index the patterns based on their frequency in the file: a more frequent pattern will get smaller index. Thus we get an integer sequence representing the file, each term denoting a pattern index. After that, we identify the monotonically increasing, decreasing and repetitive subsequences from this sequence. The increasing and decreasing subsequences are replaced by a single integer suggested by the Zeckendorf’s theorem; those that fall apart are sent as they are. All necessary accounting information is kept. The header section of the output file contains the pattern information, and the data section contains the encoded/compressed data. In all cases, this output file serves as the encrypted form of the input file; and in many cases, as the compressed form of the input file.