Application of neural networks to microstructure analysis

• Autorzy: Pamuła T.
• Języki publikacji: EN

Abstrakty

EN

The paper presents results of experiments in applying the neural network classifier to distinguish classes of microstructures on microscopic images. The designed neuron classifier incorporated basic image preprocessing and a neural toolbox based on feed-forward backpropagation multilayer net. Optionally one could modify the training algorithms of the network, the learning rate and the coefficients of the sigmoidal transfer functions. Results confirm the utmost importance of training the network with properly classified sets of input vectors. Methods of classical pattern feature extraction were used to prepare training sets. The application of PCA to enhance the performance of microstructure classification is proposed in the concluding sections of the paper.

Słowa kluczowe

EN

image processing; neural networks; microstructures

PL

przetwarzanie obrazu; sieci neuronowe; mikrostruktury

• Wydawca: University of Silesia, Institute of Informatics, Computer Systems Department
• Czasopismo: Journal of Medical Informatics & Technologies
• Rocznik: 2000
• Tom: Vol. 1
• Strony: IT 77--82
• Opis fizyczny: Bibliogr. 12 poz., rys., tab.

Twórcy

autor

Pamuła T.

• Institute of Transport, Silesian Technical University, Katowice ul. Krasińskiego 8