A Novel Approach to Binary Neighbourhood Operations Utilising a Hardware based RBF Network

P. Gaughran (Ireland)


Machine learning, neighbourhood operations, RBFs, parallel.


In this paper we present a novel approach employing a radial basis function (RBF) network to implement binary neighbourhood operations such as dilation, erosion, opening and closing. The network used is a parallel electronic hardware device (ZISC Neuroboard 576) designed by IBM Microelectronics Division. Despite increases in serial processor speed, neighbourhood operations have proven to be computationally intensive, but by utilising a multi-purpose, generalised parallel device computation overheads can be significantly reduced.

Important Links:

Go Back