Prev: How to plot (+/-) and (-/+) transitions of 3D cube matrix?
Next: motion detection
From: Datta Prasad on 31 Mar 2010 14:57
Hiiii
I am new to NN. I want to train a backpropagtion NN with 30 distinct binary imges images of 40 by 40 size.I am using nnprtool.

P is 1600 by 30 matrix containing input image data
T is 30 by 30 matrix containing output data.

Now I train with NN with 30 neurons.
and input as follows
p=[P,P,P,P,P,P,P,P,P,P];

and tragets as
t=[T,T,T,T,T,T,T,T,T,T] ;

with samples divised as 70% training 15% validation 15% testing.
Network gets trained at 34 th iteration.

==>But I am not able to see confusion plot properly.
==>And results are also not at all accurate.
Please guide me about what parameters should I vary and in what range?
Thank You
 | 
Pages: 1
Prev: How to plot (+/-) and (-/+) transitions of 3D cube matrix?
Next: motion detection