Hi, I am trying to use a neural network to recognize all upppercase letters of the alphabet, the 10 numerical digits (0-9) and neglect any other symbol not related to the above sets from a 2d image.
I guess the questions that I want to ask are: 1. How do I determine the # of initial and hidden units per layer to use. 2. Are there any resources (documents, websites, etc) out there that can aid my work 3. I read somewhere about using a sinusoidal activation function as opposed to a sigmoid one, what is a sinusoidal activation function and how do I implement it Thx in advance for any help that are given, Scott :) =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA2D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
