A Radical-Partitioned Neural Network System Using a Modified Sigmoid Function and a Wight-Dotted Radical Selector for Large-Volume Chinese Characters Recognition VLSI.
James B. KuoB. Y. ChenMark W. MaoPublished in: ISCAS (1994)
Keyphrases
- feature space
- sigmoid function
- chinese characters
- neural network
- chinese character recognition
- character recognition
- hand written
- feature extraction
- imprecise data
- activation function
- pattern recognition
- classification accuracy
- feature vectors
- trained neural network
- support vector
- linear combination
- training algorithm
- computer vision
- genetic algorithm