| Science Forum Index » Image Processing Forum » Object recognition... |
|
Page 1 of 1 |
|
| Author |
Message |
| taurus... |
 Posted: Thu Sep 17, 2009 5:33 am |
|
|
|
Guest
|
I have 2 grayscale image with the known position of 2 people in it.
How can run something on image 1 to get out features for each persson
and then compare them with image 2 people?
Like what technique can i use to compare the people? |
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Fri Sep 18, 2009 5:10 pm |
|
|
|
Guest
|
| im programming it myself. |
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Fri Sep 18, 2009 6:15 pm |
|
|
|
Guest
|
and its hard for face recognition because its a gray scale image and
the faces are hard to see. |
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Thu Sep 24, 2009 10:46 am |
|
|
|
Guest
|
| How would you use a histogram to compare objects? |
|
|
| Back to top |
|
|
|
| ImageAnalyst... |
| Posted: Thu Sep 24, 2009 12:48 pm |
|
|
|
Guest
|
On Sep 24, 4:46 pm, taurus <truetau... at (no spam) gmail.com> wrote:
[quote:08f86346c9]How would you use a histogram to compare objects?
-----------------------------------------------------------------------------------------------------------------[/quote:08f86346c9]
I hope you figured out by now that face recognition is not trivial,
due to to many factors to list here. I'm sure there have been many
Ph.D. dissertations on it.
A histogram lets you see the probability distribution function of the
intensities in the image. Obviously this is not sufficient to do
facial recognition on except in the most trivial of cases (e.g. two
subjects in database, one Caucasian and one African).
Obviously it's too complicated to do it for you in a few lines of code
here. Why don't you start your research here:
http://iris.usc.edu/Vision-Notes/bibliography/contentspeople.html#Face%20Recognition,%20Detection,%20Tracking,%20Gesture%20Recognition,%20Fingerprints,%20Biometrics
Pick a paper that sounds interesting to you and then try to do what
they did. |
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Thu Sep 24, 2009 3:26 pm |
|
|
|
Guest
|
well its not facial recognition, its whole body. So i was thinking get
the average intensity for the body and can compare with others. |
|
|
| Back to top |
|
|
|
| ImageAnalyst... |
| Posted: Thu Sep 24, 2009 3:48 pm |
|
|
|
Guest
|
On Sep 24, 9:26 pm, taurus <truetau... at (no spam) gmail.com> wrote:
[quote:65bc40c872]well its not facial recognition, its whole body. So i was thinking get
the average intensity for the body and can compare with others.
[/quote:65bc40c872]
------------------------------------------------------------------
This method will be completely unacceptable in the real world.
However, for a very simple class project (high school or undergrad at
best), and a very limited set of images it could work. Just don't
claim that success with your limited small test proves that it is a
robust method - because it's not - not by a long shot. |
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Thu Sep 24, 2009 8:59 pm |
|
|
|
Guest
|
| any other idea that would be better? |
|
|
| Back to top |
|
|
|
| ImageAnalyst... |
| Posted: Fri Sep 25, 2009 2:09 am |
|
|
|
Guest
|
|
| Back to top |
|
|
|
| taurus... |
| Posted: Thu Oct 01, 2009 5:40 am |
|
|
|
Guest
|
I read about statistical pattern recognition but not sure how to
perform it to compare images |
|
|
| Back to top |
|
|
|
| Giff... |
| Posted: Fri Oct 02, 2009 3:17 am |
|
|
|
Guest
|
taurus wrote:
[quote:9098c068ef]I read about statistical pattern recognition but not sure how to
perform it to compare images
[/quote:9098c068ef]
I am afraid you'll need to read a bit more. |
|
|
| Back to top |
|
|
|
|
