Abstract
Abstract
This is the second of two papers investigating the performance of general-purpose feature detection techniques. The first paper describes the development of a methodology to synthesize possible general feature detection face sets. Six algorithms resulting from the synthesis have been designed and implemented on a SUN Workstation in C++ using ACIS as the geometric modelling system. In this paper, extensive tests and comparative analysis are conducted on the feature detection algorithms, using carefully selected components from the public domain, mostly from the National Design Repository. The results show that the new and enhanced algorithms identify face sets that previously published algorithms cannot detect. The tests also show that each algorithm can detect, among other types, a certain type of feature that is unique to it. Hence, most of the algorithms discussed in this paper would have to be combined to obtain complete coverage.