Feature Aggregation in Iconic Model Evaluation
K. Brisdon, G. D. Sullivan, K. D. Baker · 1988
This paper presents recent work on iconic model-matching. The idea of iconic feature evaluation is reviewed, and methods for setting adaptive noise thresholds for use in feature combination are described. Extensions to the adaptive thresholding technique are explained and illustrated, and the relevance of this technique to feature combination is discussed. Finally demonstrations of the performance of the system are shown, with particular reference to the discrimination ability of the method with multiple models. This paper describes a method of model-matching, applicable as a verification procedure within a knowledge-based vision systems containing three-dimensional geometric models. Most approaches to object verification in model-based vision merely extend