Improved neocognitron with bend-detecting cells
- 2 January 2003
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 4, 190-195
- https://doi.org/10.1109/ijcnn.1992.227343
Abstract
The authors present an improved neocognitron, in which bend-detecting cells, as well as line-extracting cells, are utilized. In contrast to the conventional neocognitron, which has shown a lesser ability to recognize deformed patterns when trained using unsupervised learning than by using supervised learning, the new system shows considerable robustness even when trained by unsupervised learning. The new system can also accept some variation in the line thickness of the input patterns. Edge-extracting cells, which are built into the network, are effectively used for this purpose. A line is extracted from the edges on both sides of the line. To demonstrate the performance of the new system, the network has been trained to recognize handwritten numerals.Keywords
This publication has 5 references indexed in Scilit:
- Neocognitron: A new algorithm for pattern recognition tolerant of deformations and shifts in positionPublished by Elsevier ,2003
- Neocognitron: A hierarchical neural network capable of visual pattern recognitionPublished by Elsevier ,2003
- Handwritten alphanumeric character recognition by the neocognitronIEEE Transactions on Neural Networks, 1991
- Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in positionBiological Cybernetics, 1980
- A feature extractor for curvilinear patterns: a design suggested by the mammalian visual systemBiological Cybernetics, 1970