Институт вычислительной математики и математической геофизики СОРАН

Abstracts

The problem of effective run-length encoding for facsimile data transmission via a network is considered. For the first time, the scheme of facsimile run-length encoding was proposed by R.Hunter and A.Robinson. As an information source model, Markov chain of first order was used, the probabilities of lengths being calculated with allowance for the model statistics. A new efficient run-length encoding algorithm is proposed. In this case, Markov source of order n generating letters from alphabet A={0,1} with unknown conditional probabilities is used as a model. As distinct from the earlier known algorithms, the proposed adaptive encoding method makes it possible to attain any preassigned redundancy with conservation of a small memory size of an encoder and decoder as well as little average time of coding and decoding.

Note. Abstracts are published in author's edition

Mail to Webmaster

|Home Page| |English Part|

Go to Home