Efficient run-length encodings

Abstract
A method for the noiseless coding of binary memoryless sources when the symbol probabilities are unknown is investigated. Codes of long block length are constructed from truncated run-length codes. Though not universal, the codes are shown to achieve high efficiencies over a wide range of symbol probabilities. Several generalizations of the basic scheme are also presented. The schemes presented in this paper are particularly well-suited for facsimile coding.

This publication has 7 references indexed in Scilit: