Neil A H Campbell

Learn More
A pruning method is presented that is applicable to a one hidden layer classiication network, where each unit of that layer can be conceived as performing one discrimination task. The method, termed \task{based pruning", depends on selecting a subset of units that will satisfactorily perform the necessary tasks.
It is suggested in the literature 2, 1] that there is a natural pairing between the softmax activation function and the cross{entropy penalty function. We clarify a reason for this pairing and give an improved derivation of the softmax activation function. In addition, we empirically compare some penalty/activation function pairs.
Fitting a multi{layer perceptron (m.l.p.) model with an excessive number of parameters may lead to a classiier with a low apparent error rate but an unacceptably high true error rate. A method is presented for reducing the number of parameters to be estimated in the case where the data to be classiied are in fact spectra. The method is applicable in the(More)
During the leaf movements of Albizzia julibrissin Durazzini, volume changes in the motor cells of the pulvinule (tertiary pulvinus) are closely correlated with a reversible reorganization of the vacuolar compartment. Motor cells have central vacuoles when expanded, but become multivacuolate during the time the cell volume decreases. The central vacuole(More)
Sub-pixel matching is one of the key components for image registration and image fusion. Ideally, image matching should allow for offsets in the target image, and for scaling and rotation. Offsets allow for sub-pixel shifts in the two images, while scaling is necessary when matching images from different sensors or images taken from different distances(More)
The ion microscope was used to study potassium and calcium distributions in the main motor organ of Mimosa pudica L. The cortex of the motor organ has two cell types differing in location, structure, and ion distribution. Histochemical features portrayed in ion micrographs were plainly correlated with structures seen in the light and electron microscopes.
The method of task{based pruning (Dunne, Campbell and Kiiveri, 1992) is applied to a remotely sensed image. The output of both hidden{layer units and output units are interpreted as images, which, with some sim-pliication of the weight matrices, leads to an interpretation of the action of the MLP in classifying the image into ground cover classes. In(More)
This paper reviews a Bayesian approach to classifying multi{spectral image data where the pixel labels are assumed to be spatially correlated. A Markov Random Field (MRF) model is introduced to model localized dependence, so that a label is assumed to be conditional on the labels of the neighbouring pixels only. The Multi{layer Perceptron model is extended(More)
Turgor-mediated leaf movements of the legume Samanea saman are associated with the migration of K(+) and Cl(-) between opposing sides of the motor organs (pulvini). We have investigated the pathway of this ion migration by localizing K(+) and Cl(-) within the secondary pulvinus at various times during leaf movements. Ion distributions in freeze-dried(More)