This paper develops a new technology on image recognition of objects using neural networks. The standard method of pattern recognition is geometrical pattern matching. However, it is generally difficult to recognize figures on semitransparent glass because the brightness differences for those objects are pretty little. In this research, a new image recognition system is proposed by the application of neural networks. This is a neural network system which transformed a figure including patterns with noise and breaking form into a formed and well-regulated one. The recognition experiments are performed for ten categories using neural networks. Through the experiments, the validity of this system is clarified.
展开▼