i have already written my code using the neural network logic....and it works fine...
it is basically a hand written character recognition system...which accepts "bitmap" image files which i made through "windows Paint application" (from A to Z) and recognizes which character is it.
my logic is.....
open the "*.bmp" file using microsoft common dialog control....in the form then obtain the pixels using "getpixel " function...check if it is black then count no. of VBblack in the grid before which we.... divide the image into 5 * 5 grid if there is a black pixel in a grid
assign the value 1 else assign 0 for "non black pixel ".
check for all (A-Z) bmp files what matrix(0 or 1) we are getting on execution and print the character .
this code is working fine using the neural network logic discussed above, but i was asked to implement it with "genetic algorithms" a random optimization search technique , I have understand the concept of GA's already but i m unable to deduce a logic to implement on my character coding programme please help me!!!!