Removing Redundancy in Dictionary based Compression Techniques
Neha Gupta, R. J. S. Jeba Kumar, Apoorv Gupta · 2010
Many data compression schemes are developed nowadays and they are selected according to the requirements, such as fast encoding, fast decoding, a good compression performance, small amount of required memory etc. In this thesis, the basic dictionary based data compression techniques i.e. LZ77, LZ78 and LZW, have been studied to find their drawbacks, so that they can be improved further. As out of LZ77, LZ78 and LZW, the variants of LZW are widely used in a number of applications. So, the thesis is mainly oriented towards improving on LZW. Based on the study, we have tried to identify the sources of redundancy in these algorithms and have suggested a method which is a simple dictionary-pruning algorithm that removes the irrelevant entries from the dictionary every time the dictionary is out of space; to store new phrases. This ensures that the dictionary is always adaptive. Keyword: Compression, Decompression, Dictionary-based, LZW, Pruning, Performance.