Two combinatorial criteria for BWT images

Abstract

Burrows-Wheeler transform (BWT) is a block data transformation, i.e. a function on finite words. This function is used in lossless data compression algorithms and possesses interesting combinatorial properties. We study some of these properties. Namely, we prove two necessary and sufficient conditions concerning BWT images. The first one describes the words that are BWT images, while the second one explains which words can be converted to BWT images using a natural pumping procedure. Both conditions can be checked in linear time. © 2011 Springer-Verlag Berlin Heidelberg.