Wordspotter training using figure-of-merit back propagation.
Richard LippmannEric I. ChangCharles R. Jankowski Jr.Published in: ICASSP (1) (1994)
Keyphrases
- back propagation
- figure of merit
- training algorithm
- feed forward neural networks
- feedforward neural networks
- artificial neural networks
- multilayer perceptron neural network
- feed forward
- neural network
- multilayer neural network
- error back propagation
- hidden layer
- learning algorithm
- fuzzy logic
- bp neural network
- bp algorithm
- multilayer perceptron
- neural nets
- bp network
- edge detector
- activation function
- training process
- weight update
- levenberg marquardt
- multi layer perceptron
- multiresolution
- expert systems
- training phase
- online training
- recurrent neural networks