Coding with variable block maps

In this article we study a special class of sliding block maps that we call variable block maps. We characterize the subsets of finite and infinite sequences that can be obtained as the image of another subset of symbolic sequences by a variable block map. On the other way, we show that the coding process induced by such kind of block maps can be reversed, even with partial knowledge about the variable block maps, and we give an explicit construction of a canonical antecedent.