On the relationship between augmented transition network and attributed grammar

Finite state attributed grammars are known to be rich in descriptive power because semantic information is included along with the structural information. They have applied to a number of practical problems in pattern recognition. The generative power of a finite-state attributed grammar is studied. The approach adopted is to establish a relationship between this grammar and the augmented transition network, which is known to be as powerful as a Turing machine. It is established that the grammar can generate at least some of the type φ language. The potential application of this grammar is in the areas of computer vision and artificial intelligence