Accurate affine image registration using radial basis neural networks
Haldun Sarnel, Yavuz Şenol · The International Conference on Electrical Engineering/The International Conference on Electrical Engineering · 2008
Neural network-based image registration using global image features is relatively a newresearch subject and the schemes devised so far use a feedforward neural network tofind the geometrical transformation parameters. In this work, we propose to use a radialbasis function neural network instead of feedforward neural network to overcomelengthy pre-registration training stage. This modification has been tested on a typicalneural network-based registration method using discrete cosine transformation featuresin the presence of noise. The proposed scheme does not only speed up the training stageenormously, but also increases the accuracy and robustness against additive white noiseowing to the better generalization ability of the radial basis function neural networks.