Part of Advances in Neural Information Processing Systems 8 (NIPS 1995)
Alistair Ferguson, Theo Sabisch, Paul Kaye, Laurence Dixon, Hamid Bolouri
Current environmental monitoring systems assume particles to be spherical, and do not attempt to classify them. A laser-based sys(cid:173) tem developed at the University of Hertfordshire aims at classify(cid:173) ing airborne particles through the generation of two-dimensional scattering profiles. The pedormances of template matching, and two types of neural network (HyperNet and semi-linear units) are compared for image classification. The neural network approach is shown to be capable of comparable recognition pedormance, while offering a number of advantages over template matching.