Gaussian-type activation function with learnable parameters in complex-valued convolutional neural network and its application for PolSAR classification.
Yun ZhangQinglong HuaHaotian WangZhenyuan JiYong WangPublished in: Neurocomputing (2023)
Keyphrases
- activation function
- complex valued
- convolutional neural network
- neural network
- feed forward
- artificial neural networks
- multi layer perceptron
- hidden layer
- back propagation
- real valued
- training phase
- learning rate
- pattern recognition
- neural nets
- radial basis function
- support vector machine svm
- basis functions
- support vector
- training set
- network architecture
- recurrent neural networks
- rbf neural network
- supervised learning
- input space
- autoregressive
- face detection
- class labels
- feature extraction
- classification algorithm
- fuzzy logic
- prior knowledge
- convergence rate
- learning algorithm
- genetic algorithm