Title :
Determining the network parameters of the amphibian vision system
Author :
Brauer, E.J. ; Nishikawa, K.
Author_Institution :
Northern Arizona Univ., Flagstaff, AZ, USA
Abstract :
The objective of this research is to develop vision systems based on visual predators of the amphibian family (toads, frogs and salamanders). The initial phase of the project is modeling a retina (R2, R3, R4) and several neuron layers (T3, and TH-1), and their inhibitory and excitatory synapses using Genesis, a general purpose neural simulator. Since the network parameters are unknown, we are using genetic algorithms to search for a combination of parameters that produces the desired output of correctly identifying prey.
Keywords :
eye; genetic algorithms; image recognition; neural nets; parameter estimation; vision; amphibian family visual predators; amphibian vision system network parameters; excitatory synapses; frogs; genetic algorithms; inhibitory synapses; neural simulator; neuron layer modeling; prey identification; retina modeling; salamanders; toads; Animal structures; Computational modeling; Computer worms; Fires; Genetic algorithms; Machine vision; Neurons; Retina; Shape; Visual system;
Conference_Titel :
Circuits and Systems, 2002. MWSCAS-2002. The 2002 45th Midwest Symposium on
Print_ISBN :
0-7803-7523-8
DOI :
10.1109/MWSCAS.2002.1187244
