An algorithm for evaluation of white-black images characteristic vector

• Autorzy: Lisiecka-Frąszczak, J.
• Konferencja: Poznan Telecommunications Workshop / sympozjum [11; December 7-8, 2006; Poznan]
• Języki publikacji: EN

Abstrakty

EN

In the paper a new algorithm for the evaluation of the characteristic vector b of a given while-black image is presented. This vector is used in the perceptron synthesis to define its initial weights. The idea of vector b is connected with some Boolean functions determined by Karnaugh table. The white-black image is defined by pixels (p) and its values w(p). And so it is necessary to transform the primary definition of the vector b to a new concept exploiting the ideas determining the image: ρ and w(ρ).

Słowa kluczowe

EN

perceptron synthesis; white-black image vector characteristic

• Czasopismo: Poznan University of Technology Academic Journals. Electrical Engineering
• Rocznik: 2007
• Tom: No 54
• Strony: 125-131
• Opis fizyczny: Bibliogr. 4 poz.

Twórcy

autor

Lisiecka-Frąszczak, J.

• Poznan University of Technology

Bibliografia

• [1]M.L.Dertouzos, Threshold Logic: A Synthesis Approach, MIT Press, Cambridge, 1965.
• [2]Lisiecka-Frąszczak J., Wstępna ocena parametrów dyskretnej sieci neuronowej, KOWBAN'99, Poland, pp.363-368.
• [3]Lisiecka-Frąszczak J., The Optimal Laerning Set for the Binary Discrete Perceptron, IWSSIP02 Proceedings, Manchester, UK 2002, pp.435-462.
• [4]Lisiecka-Frąszczak J., On the Classification Method of Lineary Nonseparable Binary Images, IWSSIP03 Proceedings, Prague, Czech Republic 2003, pp. 186-189.