Title :
Minimal optimal topologies for invariant higher-order neural architectures using genetic algorithms
Author :
Liatsis, Panagiotis ; Goulermas, Yannis J P
Author_Institution :
Control Syst. Centre, Univ. of Manchester Inst. of Sci. & Technol., UK
Abstract :
Higher-order neural networks (HONNs) are successful in performing position, rotation and scale (PRSI) recognition. A major limitation of these networks is the combinatorial explosion of the higher-order terms, which increases the complexity of the network architecture. This work proposes a genetic optimisation scheme for determining the minimal optimal topology of a network for automated inspection of industrial parts
Keywords :
automatic optical inspection; computer vision; genetic algorithms; neural net architecture; neural nets; object recognition; automated inspection; coarse coding; genetic algorithms; genetic optimisation scheme; industrial parts; invariant higher-order neural architectures; minimal optimal topologies; object recognition; position recognition; rotation recognition; scale recognition; Biological systems; Data mining; Explosions; Feature extraction; Genetic algorithms; Information geometry; Machine vision; Network topology; Neural networks; Neurons;
Conference_Titel :
Industrial Electronics, 1995. ISIE '95., Proceedings of the IEEE International Symposium on
Conference_Location :
Athens
Print_ISBN :
0-7803-7369-3
DOI :
10.1109/ISIE.1995.497287
