I need to solve image recognition problem using Neural-network (backpropagation and quickprop algorithm). The image size will be 320x200 pixel (but you may read smaller part so the input won't be so large). There are 5 images to be recognized. See attached file.
This must be done using backpropagation and quickprop algorithm!
The attached file consists of 4 images to be recognized by the application.
Basically the application will have 2 kinds of process,
1. Learn (building correct weight for each layer)
2. Test (determine the output from certain input)