Prev: Title for each bar in function 'bar'
Next: implement may own activation function in neural network
From: mops zaki on 24 May 2010 09:05
ImageAnalyst <imageanalyst(a)mailinator.com> wrote in message <bd56f0b7-a182-4e11-9423-f1c9f4a994e8(a)e28g2000vbd.googlegroups.com>...
> What mean image? You have the mean value of ONE image over all pixels
> in that one image, which is a scalar. What got reduced? You still
> have your original image at its original size. So just multiply the
> scalar by the array and you get what you want. WHY you want this
> baffles me. What use could this possibly have?
>
> And why do you say you want to multiply the "mean image" by the
> "original one"? You DON"T HAVE a mean image unless you take the mean
> of SEVERAL images. And again, what possible use could that have?


i want to do normalization of an image.....after calculating mean and variance i have to subract (sorry not multyply) that image with original image.to get normalized image...
After subrataction result is in 1x1 matrix....and it show nothing in output...
may b im nt explaning u well
regardz
zaki
From: David Young on 24 May 2010 09:16
"mops zaki" <zaki_achi(a)hotmail.com> wrote in message <htdtig$3jr$1(a)fred.mathworks.com>...
> ...
> i want to do normalization of an image.....after calculating mean and variance i have to subract (sorry not multyply) that image with original image.to get normalized image...
> After subrataction result is in 1x1 matrix....and it show nothing in output...

If you subtract a scalar (that is, a 1x1 matrix) from an array, the result is the same size as the array. So if you do

image - mean(image(:))

you will have something the size of the original image. You must be subtracting one scalar from another if what you end up with is a scalar.
First  |  Prev  | 
Pages: 1 2
Prev: Title for each bar in function 'bar'
Next: implement may own activation function in neural network