Prev: Parallel Port can't get value
Next: trying to understand the kmenas result
From: shilpa khinvasara on 8 Jun 2010 02:12
how to quantize an HSV image into HUE(8 BINS),sATURATION(3 BINS),VALUE(3 BINS)
using Matlab

EMail:shipkhinvasara(a)gmail.com
From: Walter Roberson on 8 Jun 2010 08:46
shilpa khinvasara wrote:
> how to quantize an HSV image into HUE(8 BINS),sATURATION(3 BINS),VALUE(3
> BINS)
> using Matlab

Is "BINS" the same as "bits" ? If so then you have included a total of
14 bits: is that what you wanted? What output format do you need, and
how _exactly_ is the data to be packed into that output?
From: shilpa khinvasara on 9 Jun 2010 01:12
Sir,
I AM WORKING ON A PROJECT BASED ON CONTENT BASED IMAGE RETRIEVAL
i have an RGB image which is converted to HSV .I want to quantize the image into 24 bins
of which Hue(18 bins) saturation 3 bins value 3 bins.The coding i am trying to do in Matlab


Walter Roberson <roberson(a)hushmail.com> wrote in message <YYqPn.23490$%u7.18375(a)newsfe14.iad>...
> shilpa khinvasara wrote:
> > how to quantize an HSV image into HUE(8 BINS),sATURATION(3 BINS),VALUE(3
> > BINS)
> > using Matlab
>
> Is "BINS" the same as "bits" ? If so then you have included a total of
> 14 bits: is that what you wanted? What output format do you need, and
> how _exactly_ is the data to be packed into that output?
From: Walter Roberson on 9 Jun 2010 01:48
shilpa khinvasara wrote:

> I AM WORKING ON A PROJECT BASED ON CONTENT BASED IMAGE RETRIEVAL
> i have an RGB image which is converted to HSV .I want to quantize the
> image into 24 bins of which Hue(18 bins) saturation 3 bins value 3
> bins.The coding i am trying to do in Matlab

You should be able to work from this:


MaxH = 1;
if isinteger(TheImage); MaxH = double(intmax(class(TheImage))); end

Nbin = 18;
bounds = linspace(0,MaxH,Nbin+1);
bounds(end) = MaxH * (1+eps)
t = histc( reshape(TheImage(:,:,1),[],1), b);
Hquant = t(1:end-1); %the last bin is MaxH*(1+eps) exactly, no entries

repeat with different number of bins and different indices of TheImage
for S and V.
From: shilpa khinvasara on 9 Jun 2010 02:12
thank u very much.
is the code for indexed image??i am unable to understand what is 'b' in the code

shilpa
Walter Roberson <roberson(a)hushmail.com> wrote in message <AWFPn.68122$Gx2.32268(a)newsfe20.iad>...
> shilpa khinvasara wrote:
>
> > I AM WORKING ON A PROJECT BASED ON CONTENT BASED IMAGE RETRIEVAL
> > i have an RGB image which is converted to HSV .I want to quantize the
> > image into 24 bins of which Hue(18 bins) saturation 3 bins value 3
> > bins.The coding i am trying to do in Matlab
>
> You should be able to work from this:
>
>
> MaxH = 1;
> if isinteger(TheImage); MaxH = double(intmax(class(TheImage))); end
>

> Nbin = 18;
> bounds = linspace(0,MaxH,Nbin+1);
> bounds(end) = MaxH * (1+eps)
> t = histc( reshape(TheImage(:,:,1),[],1), b);
> Hquant = t(1:end-1); %the last bin is MaxH*(1+eps) exactly, no entries
>
> repeat with different number of bins and different indices of TheImage
> for S and V.
 |  Next  |  Last
Pages: 1 2
Prev: Parallel Port can't get value
Next: trying to understand the kmenas result