<![CDATA[First Process is found in who it is handwritten on the paper. The handwritten is scanning to be image digital so the its to easy feature extraction, the proces s name pre-processing. Result of the pre-processing to be feature extraction. The feature extraction is neuron input who organize by matrix. The input neuron to input inclassification neural network. Neural network who its need back propagation. Step by step who description in the top is training or execution network.
Application testing in training network with a lot of sample character, then network who have training to be execution with difference input character. The result, the network can recognition a lot of pattern character but also a lot of failed for the other. Grade of goal in network to be pattern defend in much sample representative in training, technique feature and source. ]]>
Copyrights © 2010
