hello,
I try to initialize a neural network without a hiddenlayer to estimate the distance from a wall with to circles on it. the neural network should be as simple as possible ( I don´t need to consider of a rotation or angle deviation). The inputneurons are the pixels of the images which i transform in a vector, so that it isn´t a matrix anymore. The network should connect directly between 100 inputneurons and just 1 outputeuron.
I hope you can help me with my issue.
Thanks for your help!
Best Answer