Using neural networks to learn shape decomposition by successive prototypication
I describe a neural-network which decomposes a set of inputs into a sequence of generative parameters. It uses a series of coupled parameter finding and removing networks and requires the input to be in a particular temporal format.
Unable to display preview. Download preview PDF.
- 1.Kohonen T. (1984). "Self organisation and Associative memory". Springer Verlag, Berlin 1984.Google Scholar
- 2.Saund E. (1989). "Dimensionality Reduction using Connectionist Networks". IEEE PAMI Vol 11 No 3, pp 304–314.Google Scholar
- 3.Leyton M. (1985). "Generative Systems of Analysers". Computer Vision, Graphics and Image Processing 31, pp 201–241.Google Scholar
- 4.Zipser D. and Anderson R.A. (1988). "A back-Propagation Programmed Network that Simulates Response Properties of a Subset of Posterior Parietal Neurons". Nature Vol 331, pp 679–684.Google Scholar