Forums > Jitter

New version of cv.jit available

October 4, 2006 | 2:16 am

(Apologies for cross-postings)

It took longer than I thought but the new cv.jit is out. On the Mac, Mach-O support is finally implemented so that there should be no problem using the externals with Jitter 1.6. A good number of externals have also been added to the distribution. Highlights include:

-Face detection: You can now easily identify rectangular areas likely to contain frontal views of human faces.

-Much improved labeling support: You can now compute centroids, moments and shape parameters for each blob separately. You can perform shape recognition on blobs, and a tool for re-ordering labels so that they don’t easily change from frame to frame is provided.

-Feature detection: Those who have used cv.jit.track may have noticed that some pixels "stick" better than others. You can now easily identify such pixels beforehand and initialize tracking to the best possible position.

-Line detection: Find straight lines in greyscale images.

-Drawing tools: Several abstractions are available for displaying data on incoming images.

-Though the pdf manual is no longer updated, every external and abstraction (with the exception of the drawing tools) have their own help files.

-Some objects have been deprecated but are still provided for patch compatibility.

http://www.iamas.ac.jp/~jovan02/cv/

Jean-Marc Pelletier


October 4, 2006 | 3:15 am

Awesome! The new additions are really amazing. Nice work.

wes

On 10/3/06, Jean-Marc Pelletier wrote:
>
> (Apologies for cross-postings)
>
> It took longer than I thought but the new cv.jit is out. On the Mac, Mach-O support is finally implemented so that there should be no problem using the externals with Jitter 1.6. A good number of externals have also been added to the distribution. Highlights include:
>
> -Face detection: You can now easily identify rectangular areas likely to contain frontal views of human faces.
>
> -Much improved labeling support: You can now compute centroids, moments and shape parameters for each blob separately. You can perform shape recognition on blobs, and a tool for re-ordering labels so that they don’t easily change from frame to frame is provided.
>
> -Feature detection: Those who have used cv.jit.track may have noticed that some pixels "stick" better than others. You can now easily identify such pixels beforehand and initialize tracking to the best possible position.
>
> -Line detection: Find straight lines in greyscale images.
>
> -Drawing tools: Several abstractions are available for displaying data on incoming images.
>
> -Though the pdf manual is no longer updated, every external and abstraction (with the exception of the drawing tools) have their own help files.
>
> -Some objects have been deprecated but are still provided for patch compatibility.
>
> http://www.iamas.ac.jp/~jovan02/cv/
>
>
> Jean-Marc Pelletier
>
>
>


October 4, 2006 | 4:25 am

happy is this day

2006/10/4, Wesley Smith :
>
> Awesome! The new additions are really amazing. Nice work.
>
> wes
>
> On 10/3/06, Jean-Marc Pelletier wrote:
> >
> > (Apologies for cross-postings)
> >
> > It took longer than I thought but the new cv.jit is out. On the Mac,
> Mach-O support is finally implemented so that there should be no problem
> using the externals with Jitter 1.6. A good number of externals have also
> been added to the distribution. Highlights include:
> >
> > -Face detection: You can now easily identify rectangular areas likely to
> contain frontal views of human faces.
> >
> > -Much improved labeling support: You can now compute centroids, moments
> and shape parameters for each blob separately. You can perform shape
> recognition on blobs, and a tool for re-ordering labels so that they don’t
> easily change from frame to frame is provided.
> >
> > -Feature detection: Those who have used cv.jit.track may have noticed
> that some pixels "stick" better than others. You can now easily identify
> such pixels beforehand and initialize tracking to the best possible
> position.
> >
> > -Line detection: Find straight lines in greyscale images.
> >
> > -Drawing tools: Several abstractions are available for displaying data
> on incoming images.
> >
> > -Though the pdf manual is no longer updated, every external and
> abstraction (with the exception of the drawing tools) have their own help
> files.
> >
> > -Some objects have been deprecated but are still provided for patch
> compatibility.
> >
> > http://www.iamas.ac.jp/~jovan02/cv/
> >
> >
> > Jean-Marc Pelletier
> >
> >
> >
>


October 4, 2006 | 4:58 am

the win 1.5 zip file is corrupted, anybody else?

2006/10/4, yair reshef < yair99@gmail.com>:
>
> happy is this day
>
> 2006/10/4, Wesley Smith < wesley.hoke@gmail.com>:
> >
> > Awesome! The new additions are really amazing. Nice work.
> >
> > wes
> >
> > On 10/3/06, Jean-Marc Pelletier < gustave433@yahoo.fr> wrote:
> > >
> > > (Apologies for cross-postings)
> > >
> > > It took longer than I thought but the new cv.jit is out. On the Mac,
> > Mach-O support is finally implemented so that there should be no problem
> > using the externals with Jitter 1.6. A good number of externals have
> > also been added to the distribution. Highlights include:
> > >
> > > -Face detection: You can now easily identify rectangular areas likely
> > to contain frontal views of human faces.
> > >
> > > -Much improved labeling support: You can now compute centroids,
> > moments and shape parameters for each blob separately. You can perform shape
> > recognition on blobs, and a tool for re-ordering labels so that they don’t
> > easily change from frame to frame is provided.
> > >
> > > -Feature detection: Those who have used cv.jit.track may have noticed
> > that some pixels "stick" better than others. You can now easily identify
> > such pixels beforehand and initialize tracking to the best possible
> > position.
> > >
> > > -Line detection: Find straight lines in greyscale images.
> > >
> > > -Drawing tools: Several abstractions are available for displaying data
> > on incoming images.
> > >
> > > -Though the pdf manual is no longer updated, every external and
> > abstraction (with the exception of the drawing tools) have their own help
> > files.
> > >
> > > -Some objects have been deprecated but are still provided for patch
> > compatibility.
> > >
> > > http://www.iamas.ac.jp/~jovan02/cv/< http://www.iamas.ac.jp/%7Ejovan02/cv/>
> > >
> > >
> > > Jean-Marc Pelletier
> > >
> > >
> > >
> >
>
>


October 4, 2006 | 8:16 am

> the win 1.5 zip file is corrupted, anybody else?]

Yikes! It is. I uploaded it again, it seems fine now.


October 4, 2006 | 12:40 pm

Jean-mark, the cv.jit.blobs.sort is worth its weight in gold (how much does
an object weigh?)
until now i used eyesweb via udp to track my blobs. this is no longer
necessary.
in my book cv.jit is now the definitive lib for motion tracking in jitter.
i am sure many beautifully projects will begin with this simplification of
the tracking process. thank you again and again

2006/10/4, Jean-Marc Pelletier :
>
>
>
> > the win 1.5 zip file is corrupted, anybody else?]
>
>
> Yikes! It is. I uploaded it again, it seems fine now.
>
>


Viewing 6 posts - 1 through 6 (of 6 total)