100 likes | 134 Views
Pattern Recognition Using Artificial Neural Networks. Ace of Hearts. Eyal Ittah (60407301) Ittai Doron (53084489). Introduction. Classify images of playing cards by their suit and value. Pattern Recognition. Sensor – existing images
E N D
Pattern Recognition Using Artificial Neural Networks Ace of Hearts Eyal Ittah (60407301) Ittai Doron (53084489)
Introduction • Classify images of playing cards by their suit and value
Pattern Recognition • Sensor – existing images • Feature extraction mechanism – Relaxation Labeling on color images • Classification scheme –Artificial Neural Network used to classify suit and value
Relaxation Labeling • Objects – pixels • Labels – object, background • Initial confidence – degree of white
Artificial Neural Networks • Neurons as basic calculation units • Input layer - Each neuron a single input variable • Hidden layers • Output layer - Each neuron represents a single output variable.
Artificial Neural Networks • Training the network • Supervised learning • Back-propagation • Learning rate
Conclusions • Using two ANNs instead of one • Decreasing the learning rate of the ANN • Using an output neuron for each class