Classification of images using spatial random trees
W. Wang, Ilya Pollak, Charles A. Bouman, Mary P. Harper · IEEE/SP 13th Workshop on Statistical Signal Processing, 2005 · 2005
In this paper a new methodology for constructing hierarchical stochastic image models called spatial random trees (SRTs) is described. The models are stochastic processes on trees with each leaf corresponding to a single pixel. The key innovation is that the tree structure itself is random and is generated by a probabilistic grammar. Important application of this method is its wide usage in natural language processing