AN INEXACT GRAPH MATCHING ALGORITHM FOR THREE-DIMENSIONAL OBJECT RECOGNITION
Hans Peter Amann, Heinz Hügli · Series in machine perception and artificial intelligence · 1993
This paper describes in detail the inexact matching algorithm which has been applied with success to 3-D object representations in a 3-D object recognition system for range data. The algorithm combines in a promising way several approaches proposed in the last couple of years: an extension of the search strategy to cope with inexact matching of attributed sub-graphs, the determination of a measure for attributal and structural dissimilarities, and the global transformation fitting, features which are used to perform efficient search-tree pruning. The algorithm was tested successfully in a series of experiments involving scenes with single and multiple objects. 1.