An arithmetic coding is well known as a powerful and an interesting technique from information theoretical viewpoints. For the practical use, however, some fundamental problems such as an error propagation and an adaptive scheme are remained to be settled. In order to apply this technique for the areas of image data compression, a simplified method for the computation of augend have to be developed.
In this paper, we present a feasible method by an introduction of a dummy symbol. Since we choose an occurrence probability of dummy symbol to be very small, we can avoid the loss of a total performance. A simple simulation study is carried out for the illustration of above method developed here.
View full abstract