troikas

i would like to do pattern recognition with jitter. the

aim is to let the shape of a bottle be recognized by "cv.jit.faces ", but i

hello everybody, 
i would like to do pattern recognition with jitter. the
aim is to let the shape of a bottle be recognized by "cv.jit.faces ", but i


cv-jit-faces-different-shapes

cv.jit.faces is based on OpenCV library, so you can find the info for 

http://opencvlibrary.sourceforge.net/ObjectDetection

pdf file explaining how to create xmls(this guy created files for detecting 

cv.jit.faces is based on OpenCV library, so you can find the info for 
creating xml files here:

On that page is a link to http://robotik.inflomatik.info/ where you can find 
pdf file explaining how to create xmls(this guy created files for detecting 
cups), and windows app for training.

hey, great, thanks a lot!! this really helps.

http://www.intel.com/technology/itj/2005/volume09issue02/art03_learning_vision/p04_face_detection.htm

http://cs.uccs.edu/~jspothar/haartraining.htm

http://robotik.inflomatik.info/other/opencv/OpenCV_ObjectDetection_HowTo.pdf

You'll need an app called haarconv to convert the training data to xml.

http://tech.groups.yahoo.com/group/OpenCV/files/

You'll need to register to the OpenCV group before you can download it, though.

Some more reading:
http://www.intel.com/technology/itj/2005/volume09issue02/art03_learning_vision/p04_face_detection.htm

You'll need to register to the OpenCV group before you can download it, though.


I guess this is a pretty basic problem but don't find the answer.

I have an installation in wich there is some live camera input, when it recognizes a plain frontal face it should scale it to fullscreen and make a picture out of which then be used in another patch.

So I need to find the area which inncludes the face, I guess this is about the green square from within the cv.jit.face help patch and resize the original picture (which is of course bigger then the tracking one) to that size.

I'm pretty new to CV.jit library and have no real idea where to start.

I have an installation in wich there is some live camera input, when it recognizes a plain frontal face it should scale it to fullscreen and make a picture out of which then be used in another patch. 

The output of cv.jit.faces is a 4-plane matrix, where each cell contains the coordinates of the area that contains a face. You can use jit.iter to get the data. You can then use the output with jit.submatrix to do what you want.

Quote: Jean-Marc Pelletier wrote on Sun, 07 January 2007 23:09

----------------------------------------------------

> The output of cv.jit.faces is a 4-plane matrix, where each cell contains the coordinates of the area that contains a face. You can use jit.iter to get the data. You can then use the output with jit.submatrix to do what you want.

I am trying to get the coordinates from cv.jit.faces as well, but how do you use jit.iter to get the data?

The only way I know now to get the data is to use 'getcell' on a matrix, but I think I am missing something..

Quote: Jean-Marc Pelletier wrote on Sun, 07 January 2007 23:09
----------------------------------------------------
> The output of cv.jit.faces is a 4-plane matrix, where each cell contains the coordinates of the area that contains a face. You can use jit.iter to get the data. You can then use the output with jit.submatrix to do what you want.
>
----------------------------------------------------

I am trying to get the coordinates from cv.jit.faces as well, but how do you use jit.iter to get the data?
The only way I know now to get the data is to use 'getcell' on a matrix, but I think I am missing something..

Quote: Daan wrote on Tue, 06 March 2007 08:46

> I am trying to get the coordinates from cv.jit.faces as well, but how do you use jit.iter to get the data?

> The only way I know now to get the data is to use 'getcell' on a matrix, but I think I am missing something..

> I am trying to get the coordinates from cv.jit.faces as well, but how do you use jit.iter to get the data?
> The only way I know now to get the data is to use 'getcell' on a matrix, but I think I am missing something..

Jitter tutorial pp. 101-102.
jit.iter help file.

Is there any information if I am using a mac?

Quote: daniel.goggin@gmail.com wrote on Thu, 08 March 2007 20:58

> Is there any information if I am using a mac?

Quote: daniel.goggin@gmail.com wrote on Thu, 08 March 2007 20:58
----------------------------------------------------
> Is there any information if I am using a mac?

Are there any useful tutorials regarding making usable .xml files for reading into cv.jit.faces if one were using a mac.

Quote: daniel.goggin@gmail.com wrote on Thu, 08 March 2007 23:22

> Are there any useful tutorials regarding making usable .xml files for reading into cv.jit.faces if one were using a mac.

The latest version of OpenCV works on OS X. You could try compiling the Haartraining app yourself. I haven't done so yet on Mac.

Quote: daniel.goggin@gmail.com wrote on Thu, 08 March 2007 23:22
----------------------------------------------------
> Clarification:
>
> Are there any useful tutorials regarding making usable .xml files for reading into cv.jit.faces if one were using a mac.

The latest version of OpenCV works on OS X. You could try compiling the Haartraining app yourself. I haven't done so yet on Mac.


cv.jit.faces - different shapes?