Automatic design of W-operators using LVQ - application to morphological image segmentation
Franklin César Flores, Sarajane Marques Peres, Fernando José Von Zuben · 2003
W-operators compose a family of translation invariant and locally defined operators. This work proposes a new approach to an automatic design of W-operators using learning vector quantization neural networks. It is also presented here some experimental results associated with the application of these operators in morphological image segmentation.