Creating Ethovision style tracking with ImageJ

[Jeff-117]

Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the program
for Ethovision-like tracking. For those familiar with both programs, is this
possible? I have tens of hours of continuous video of a white rat moving
about on a black/gray background, and I would like a program that will
automatically track it. Because of the lighting, in some areas the rat
appears quite large and bright, and other times rather dim. At all times,
however, the contrast of the rat to the background is relatively high. I
have checked out several of the object/particle tracking plug-ins, and I'm
not quite certain if they will do what I need, mostly because of the
"automatic" requirement.

Ethovision is a good program for doing this sort of thing, but it is simply
too touchy when actually identifying the rat in poor or variable lighting
conditions. So, does anyone think that ImageJ can provide a solution? If so,
how?

Thanks to you all,
Jeff

[Jonathan Andrade Silva]

Hi Jeff
I am working in a program for analyzes comportamental of animals utilizing
the functionalities of the imagej that system is the system Topolino, that
tools analyzes the behavior of mice and already is in phase of conlcusão.

it visits our page: http://www.ec.ucdb.br/~pistori/GPEC/
------------------------------------------------------
Jonathan de Andrade Silva
Brazil - Mato Grosso do Sul - Campo Grande
UCDB (University Catholic Dom Bosco)
GPEC (Computation Engineering Research Group)
---------------------------------------------------------


2006/7/12, Jeff <[hidden email]>:

>
> Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the program
> for Ethovision-like tracking. For those familiar with both programs, is
> this
> possible? I have tens of hours of continuous video of a white rat moving
> about on a black/gray background, and I would like a program that will
> automatically track it. Because of the lighting, in some areas the rat
> appears quite large and bright, and other times rather dim. At all times,
> however, the contrast of the rat to the background is relatively high. I
> have checked out several of the object/particle tracking plug-ins, and I'm
> not quite certain if they will do what I need, mostly because of the
> "automatic" requirement.
>
> Ethovision is a good program for doing this sort of thing, but it is
> simply
> too touchy when actually identifying the rat in poor or variable lighting
> conditions. So, does anyone think that ImageJ can provide a solution? If
> so,
> how?
>
> Thanks to you all,
> Jeff
>

[Jonathan Andrade Silva]

Hi jeff
If you be not imported of us send your video.

2006/7/12, Jeff <[hidden email]>:

>
> Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the program
> for Ethovision-like tracking. For those familiar with both programs, is
> this
> possible? I have tens of hours of continuous video of a white rat moving
> about on a black/gray background, and I would like a program that will
> automatically track it. Because of the lighting, in some areas the rat
> appears quite large and bright, and other times rather dim. At all times,
> however, the contrast of the rat to the background is relatively high. I
> have checked out several of the object/particle tracking plug-ins, and I'm
> not quite certain if they will do what I need, mostly because of the
> "automatic" requirement.
>
> Ethovision is a good program for doing this sort of thing, but it is
> simply
> too touchy when actually identifying the rat in poor or variable lighting
> conditions. So, does anyone think that ImageJ can provide a solution? If
> so,
> how?
>
> Thanks to you all,
> Jeff
>

[Jonathan Andrade Silva]

Hi Jeff
I am working in a program for analyzes comportamental of animals utilizing
the functionalities of the imagej that system is the system Topolino, that
tools analyzes the behavior of mice and already is in phase of conlcusão.

------------------------------------------------------
Jonathan de Andrade Silva
Brazil - Mato Grosso do Sul - Campo Grande
UCDB (University Catholic Dom Bosco)
GPEC (Computation Engineering Research Group)
---------------------------------------------------------


2006/7/12, Jeff <[hidden email]>:

>
> Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the program
> for Ethovision-like tracking. For those familiar with both programs, is
> this
> possible? I have tens of hours of continuous video of a white rat moving
> about on a black/gray background, and I would like a program that will
> automatically track it. Because of the lighting, in some areas the rat
> appears quite large and bright, and other times rather dim. At all times,
> however, the contrast of the rat to the background is relatively high. I
> have checked out several of the object/particle tracking plug-ins, and I'm
> not quite certain if they will do what I need, mostly because of the
> "automatic" requirement.
>
> Ethovision is a good program for doing this sort of thing, but it is
> simply
> too touchy when actually identifying the rat in poor or variable lighting
> conditions. So, does anyone think that ImageJ can provide a solution? If
> so,
> how?
>
> Thanks to you all,
> Jeff
>

[Jeff-117]

It seems the website that you posted is down. Can you tell me some more
about your project? Are there any sites that describe it in English?

On 7/13/06, Jonathan Andrade Silva <[hidden email]> wrote:

>
> Hi Jeff
> I am working in a program for analyzes comportamental of animals utilizing
> the functionalities of the imagej that system is the system Topolino, that
> tools analyzes the behavior of mice and already is in phase of conlcusão.
>
> ------------------------------------------------------
> Jonathan de Andrade Silva
> Brazil - Mato Grosso do Sul - Campo Grande
> UCDB (University Catholic Dom Bosco)
> GPEC (Computation Engineering Research Group)
> ---------------------------------------------------------
>
>
> 2006/7/12, Jeff <[hidden email]>:
> >
> > Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the
> program
> > for Ethovision-like tracking. For those familiar with both programs, is
> > this
> > possible? I have tens of hours of continuous video of a white rat moving
> > about on a black/gray background, and I would like a program that will
> > automatically track it. Because of the lighting, in some areas the rat
> > appears quite large and bright, and other times rather dim. At all
> times,
> > however, the contrast of the rat to the background is relatively high. I
> > have checked out several of the object/particle tracking plug-ins, and
> I'm
> > not quite certain if they will do what I need, mostly because of the
> > "automatic" requirement.
> >
> > Ethovision is a good program for doing this sort of thing, but it is
> > simply
> > too touchy when actually identifying the rat in poor or variable
> lighting
> > conditions. So, does anyone think that ImageJ can provide a solution? If
> > so,
> > how?
> >
> > Thanks to you all,
> > Jeff
> >
>

[Jonathan Andrade Silva]

This project had beginning with in partnership with the course of
physiotherapy that carry out tests with drugs or genetic alterations in
mice.  For those tests exist several experiments between the which we are
working is the test of the open field.  The test of the open field consists
of an arena circulate with a wrapped up clear cylindrical one, the mouse is
put in the center of the arena after to be drugged or will suffer genetic
alterations, and then analyzes-itself his behavior, as nivel of stress of
the animal, time of permanence in determined area, the speed, distance
traversed.  The system topolino was developed for carry out those analyses
and return for the user in form of graphics the results obtained during the
experiment, I possessed also a database for store the facts of the
experiments, I possessed techniques of rastreamento and segmentation where
the user can choose to technical that better represent his results.  That
system entirely was developed in environment linux and in JAVA.

this site is in Portuguese
http://www.ec.ucdb.br/~pistori/topolino/

2006/7/13, Jeff <[hidden email]>:

>
> It seems the website that you posted is down. Can you tell me some more
> about your project? Are there any sites that describe it in English?
>
> On 7/13/06, Jonathan Andrade Silva <[hidden email]> wrote:
> >
> > Hi Jeff
> > I am working in a program for analyzes comportamental of animals
> utilizing
> > the functionalities of the imagej that system is the system Topolino,
> that
> > tools analyzes the behavior of mice and already is in phase of
> conlcusão.
> >
> > ------------------------------------------------------
> > Jonathan de Andrade Silva
> > Brazil - Mato Grosso do Sul - Campo Grande
> > UCDB (University Catholic Dom Bosco)
> > GPEC (Computation Engineering Research Group)
> > ---------------------------------------------------------
> >
> >
> > 2006/7/12, Jeff <[hidden email]>:
> > >
> > > Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the
> > program
> > > for Ethovision-like tracking. For those familiar with both programs,
> is
> > > this
> > > possible? I have tens of hours of continuous video of a white rat
> moving
> > > about on a black/gray background, and I would like a program that will
> > > automatically track it. Because of the lighting, in some areas the rat
> > > appears quite large and bright, and other times rather dim. At all
> > times,
> > > however, the contrast of the rat to the background is relatively high.
> I
> > > have checked out several of the object/particle tracking plug-ins, and
> > I'm
> > > not quite certain if they will do what I need, mostly because of the
> > > "automatic" requirement.
> > >
> > > Ethovision is a good program for doing this sort of thing, but it is
> > > simply
> > > too touchy when actually identifying the rat in poor or variable
> > lighting
> > > conditions. So, does anyone think that ImageJ can provide a solution?
> If
> > > so,
> > > how?
> > >
> > > Thanks to you all,
> > > Jeff
> > >
> >
>

[Jeff-117]

Excellent. I appreciate anything written for Linux. Is there any English
documentation anywhere?

Jeff

On 7/18/06, Jonathan Andrade Silva <[hidden email]> wrote:

>
> This project had beginning with in partnership with the course of
> physiotherapy that carry out tests with drugs or genetic alterations in
> mice.  For those tests exist several experiments between the which we are
> working is the test of the open field.  The test of the open field
> consists
> of an arena circulate with a wrapped up clear cylindrical one, the mouse
> is
> put in the center of the arena after to be drugged or will suffer genetic
> alterations, and then analyzes-itself his behavior, as nivel of stress of
> the animal, time of permanence in determined area, the speed, distance
> traversed.  The system topolino was developed for carry out those analyses
> and return for the user in form of graphics the results obtained during
> the
> experiment, I possessed also a database for store the facts of the
> experiments, I possessed techniques of rastreamento and segmentation where
> the user can choose to technical that better represent his results.  That
> system entirely was developed in environment linux and in JAVA.
>
> this site is in Portuguese
> http://www.ec.ucdb.br/~pistori/topolino/
>
> 2006/7/13, Jeff <[hidden email]>:
> >
> > It seems the website that you posted is down. Can you tell me some more
> > about your project? Are there any sites that describe it in English?
> >
> > On 7/13/06, Jonathan Andrade Silva <[hidden email]> wrote:
> > >
> > > Hi Jeff
> > > I am working in a program for analyzes comportamental of animals
> > utilizing
> > > the functionalities of the imagej that system is the system Topolino,
> > that
> > > tools analyzes the behavior of mice and already is in phase of
> > conlcusão.
> > >
> > > ------------------------------------------------------
> > > Jonathan de Andrade Silva
> > > Brazil - Mato Grosso do Sul - Campo Grande
> > > UCDB (University Catholic Dom Bosco)
> > > GPEC (Computation Engineering Research Group)
> > > ---------------------------------------------------------
> > >
> > >
> > > 2006/7/12, Jeff <[hidden email]>:
> > > >
> > > > Hello, I'm soon to be a new ImageJ user, and I'm hoping to use the
> > > program
> > > > for Ethovision-like tracking. For those familiar with both programs,
> > is
> > > > this
> > > > possible? I have tens of hours of continuous video of a white rat
> > moving
> > > > about on a black/gray background, and I would like a program that
> will
> > > > automatically track it. Because of the lighting, in some areas the
> rat
> > > > appears quite large and bright, and other times rather dim. At all
> > > times,
> > > > however, the contrast of the rat to the background is relatively
> high.
> > I
> > > > have checked out several of the object/particle tracking plug-ins,
> and
> > > I'm
> > > > not quite certain if they will do what I need, mostly because of the
> > > > "automatic" requirement.
> > > >
> > > > Ethovision is a good program for doing this sort of thing, but it is
> > > > simply
> > > > too touchy when actually identifying the rat in poor or variable
> > > lighting
> > > > conditions. So, does anyone think that ImageJ can provide a
> solution?
> > If
> > > > so,
> > > > how?
> > > >
> > > > Thanks to you all,
> > > > Jeff
> > > >
> > >
> >
>

[Pedro Alejandro Herve Fernandez Carrera Licenciatura en Cs.Biologicas]

hi there, is there a way to get TOPOLINO, i´ve done some behavior
experiments, but I used the multitracker plugin, with this you can get the
total distance travelled by the animal, and x,y coordinates.
to count the turns (left or right), I have done it just looking at each
picture (so it´s really slow), what kind of other tricks does TOPOLINO.

thank´s

Pedro A. Hervé F.
Licenciatura en Ciencias Biologicas
Universidad Austral de Chile
 

[Michael Ji]

I am using ImageJ to track  vehicle object. I look at
the source code and the problem I found is that
multitracker only works efficiently if all objects are
within frame. Am I right?

So I am modifying the code now, hope can track objects
even object is running out of the frame.

Any suggestions?

Michael,

--- "Pedro Alejandro Herve Fernandez Carrera
Licenciatura en Cs.              Biologicas"
<[hidden email]> wrote:

> hi there, is there a way to get TOPOLINO, i´ve done
> some behavior
> experiments, but I used the multitracker plugin,
> with this you can get the
> total distance travelled by the animal, and x,y
> coordinates.
> to count the turns (left or right), I have done it
> just looking at each
> picture (so it´s really slow), what kind of other
> tricks does TOPOLINO.
>
> thank´s
>
> Pedro A. Hervé F.
> Licenciatura en Ciencias Biologicas
> Universidad Austral de Chile
>  
>

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

[Nico Stuurman-4]

Check out Mtrack2
(http://valelab.ucsf.edu/~nico/IJplugins/MTrack2.html) which is based
on the Mtrack plugin and deals with the problem you describe.

Nico

On Jul 23, 2006, at 9:36 AM, Michael Ji wrote:

> I am using ImageJ to track  vehicle object. I look at
> the source code and the problem I found is that
> multitracker only works efficiently if all objects are
> within frame. Am I right?
>
> So I am modifying the code now, hope can track objects
> even object is running out of the frame.
>
> Any suggestions?
>
> Michael,
>
> --- "Pedro Alejandro Herve Fernandez Carrera
> Licenciatura en Cs.              Biologicas"
> <[hidden email]> wrote:
>
>> hi there, is there a way to get TOPOLINO, i´ve done
>> some behavior
>> experiments, but I used the multitracker plugin,
>> with this you can get the
>> total distance travelled by the animal, and x,y
>> coordinates.
>> to count the turns (left or right), I have done it
>> just looking at each
>> picture (so it´s really slow), what kind of other
>> tricks does TOPOLINO.
>>
>> thank´s
>>
>> Pedro A. Hervé F.
>> Licenciatura en Ciencias Biologicas
>> Universidad Austral de Chile
>>
>>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>
>

Vale Lab
HHMI / Dept. of Cellular and Molecular Pharmacology
University of California, San Francisco
Genentech Hall, Room N316
600 16th street

For mail:
San Francisco, CA 94143-2200

For deliveries:
San Francisco, CA 94158

email: [hidden email]
phone: [phone]
fax: [phone]

[Michael Ji]

hi Nico:

Thanks your Java package.

I look through the java code and found a bit confusion
that I am not sure.

In line 163 of MTrack2_.java is a loop
"for (int iF=i+1; iF<=(nFrames-1);iF++)"

The logic looks like that particles in the current
frame will compare with the particles in all the other
frames one by one till the end.

1.
Actually, my thought is to only compare the particles
in current frame with those in its' consequential
frame. Means, the particle identification only happens
for particles in two frames if these frames are next
each other. Is it right?

2.
The current algorithm is to figure out if two
particles are identical via their distance (shortest
distance will win). I am thinking to introduce other
property paraemters, such as, particle's area size  as
additional determining criteria.

I will go ahead to implement it. Will that be a good
idea?

thanks,

Michael,

--- Nico Stuurman <[hidden email]> wrote:

> Check out Mtrack2
>
(http://valelab.ucsf.edu/~nico/IJplugins/MTrack2.html)

> which is based
> on the Mtrack plugin and deals with the problem you
> describe.
>
> Nico
>
> On Jul 23, 2006, at 9:36 AM, Michael Ji wrote:
>
> > I am using ImageJ to track  vehicle object. I look
> at
> > the source code and the problem I found is that
> > multitracker only works efficiently if all objects
> are
> > within frame. Am I right?
> >
> > So I am modifying the code now, hope can track
> objects
> > even object is running out of the frame.
> >
> > Any suggestions?
> >
> > Michael,
> >
> > --- "Pedro Alejandro Herve Fernandez Carrera
> > Licenciatura en Cs.              Biologicas"
> > <[hidden email]> wrote:
> >
> >> hi there, is there a way to get TOPOLINO, i´ve
> done
> >> some behavior
> >> experiments, but I used the multitracker plugin,
> >> with this you can get the
> >> total distance travelled by the animal, and x,y
> >> coordinates.
> >> to count the turns (left or right), I have done
> it
> >> just looking at each
> >> picture (so it´s really slow), what kind of other
> >> tricks does TOPOLINO.
> >>
> >> thank´s
> >>
> >> Pedro A. Hervé F.
> >> Licenciatura en Ciencias Biologicas
> >> Universidad Austral de Chile
> >>
> >>
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam?  Yahoo! Mail has the best spam
> protection around
> > http://mail.yahoo.com
> >
> >
> Vale Lab
> HHMI / Dept. of Cellular and Molecular Pharmacology
> University of California, San Francisco
> Genentech Hall, Room N316
> 600 16th street
>
> For mail:
> San Francisco, CA 94143-2200
>
> For deliveries:
> San Francisco, CA 94158
>
> email: [hidden email]
> phone: [phone]
> fax: [phone]
>

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

[Nico Stuurman-4]

> 1.
> Actually, my thought is to only compare the particles
> in current frame with those in its' consequential
> frame. Means, the particle identification only happens
> for particles in two frames if these frames are next
> each other. Is it right?

The Mtrack2 run function works in two stages.  In the first stage,
particles are found in all frames using the ImageJ ParticleAnalyzer
class.  In the second stage, these particles are assembled in tracks
that span multiple frames.  Indeed, the plugin starts each track using
any given particle not part of a track yet and then tries to find
particles 'belonging' to that track in subsequent frames.  Any other
approach should work (but should give the exact same result).

> 2.
> The current algorithm is to figure out if two
> particles are identical via their distance (shortest

That is right.  The plugin was specifically developed to track
diffraction limited spots that are diffusing randomly.  Thus, there
were no other criteria (size, shape, texture, intensity, velocity,
etc..) that could reliably used to identify the particle.  That is also
the reason why the Maximum Velocity parameter is so important.  If you
are in a situation where you can use other criteria to identify a
particle between frames, then you should most certainly do so (but it
would be nice to make that optional so that the plugin remains more
widely usable)..

Nico

> distance will win). I am thinking to introduce other
> property paraemters, such as, particle's area size  as
> additional determining criteria.
>
> I will go ahead to implement it. Will that be a good
> idea?
>
> thanks,
>
> Michael,
>
> --- Nico Stuurman <[hidden email]> wrote:
>
>> Check out Mtrack2
>>
> (http://valelab.ucsf.edu/~nico/IJplugins/MTrack2.html)
>> which is based
>> on the Mtrack plugin and deals with the problem you
>> describe.
>>
>> Nico
>>
>> On Jul 23, 2006, at 9:36 AM, Michael Ji wrote:
>>
>>> I am using ImageJ to track  vehicle object. I look
>> at
>>> the source code and the problem I found is that
>>> multitracker only works efficiently if all objects
>> are
>>> within frame. Am I right?
>>>
>>> So I am modifying the code now, hope can track
>> objects
>>> even object is running out of the frame.
>>>
>>> Any suggestions?
>>>
>>> Michael,
>>>
>>> --- "Pedro Alejandro Herve Fernandez Carrera
>>> Licenciatura en Cs.              Biologicas"
>>> <[hidden email]> wrote:
>>>
>>>> hi there, is there a way to get TOPOLINO, i´ve
>> done
>>>> some behavior
>>>> experiments, but I used the multitracker plugin,
>>>> with this you can get the
>>>> total distance travelled by the animal, and x,y
>>>> coordinates.
>>>> to count the turns (left or right), I have done
>> it
>>>> just looking at each
>>>> picture (so it´s really slow), what kind of other
>>>> tricks does TOPOLINO.
>>>>
>>>> thank´s
>>>>
>>>> Pedro A. Hervé F.
>>>> Licenciatura en Ciencias Biologicas
>>>> Universidad Austral de Chile
>>>>
>>>>
>>>
>>>
>>> __________________________________________________
>>> Do You Yahoo!?
>>> Tired of spam?  Yahoo! Mail has the best spam
>> protection around
>>> http://mail.yahoo.com
>>>
>>>
>> Vale Lab
>> HHMI / Dept. of Cellular and Molecular Pharmacology
>> University of California, San Francisco
>> Genentech Hall, Room N316
>> 600 16th street
>>
>> For mail:
>> San Francisco, CA 94143-2200
>>
>> For deliveries:
>> San Francisco, CA 94158
>>
>> email: [hidden email]
>> phone: [phone]
>> fax: [phone]
>>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>
>

Vale Lab
HHMI / Dept. of Cellular and Molecular Pharmacology
University of California, San Francisco
Genentech Hall, Room N316
600 16th street

For mail:
San Francisco, CA 94143-2200

For deliveries:
San Francisco, CA 94158

email: [hidden email]
phone: [phone]
fax: [phone]

[Michael Ji]

 

> > 2.
> > The current algorithm is to figure out if two
> > particles are identical via their distance
> (shortest
>
> That is right.  The plugin was specifically
> developed to track
> diffraction limited spots that are diffusing
> randomly.  Thus, there
> were no other criteria (size, shape, texture,
> intensity, velocity,
> etc..) that could reliably used to identify the
> particle.  That is also
> the reason why the Maximum Velocity parameter is so
> important.  If you
> are in a situation where you can use other criteria
> to identify a
> particle between frames, then you should most
> certainly do so (but it
> would be nice to make that optional so that the
> plugin remains more
> widely usable)..
>
> Nico
>

thanks your comment about provide option to the user;

I met an issue that two particles are not identical,
but are viewed as same object, because they are the
most closed ones between frames. I guess it is mostly
due to my particular application --- tracking vehicles
between frames. In that case, some vehicles in the
opposite direction will confuse the algorithm.

Michael,

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

[Michael Ji]

Hi there,

I try the following codes to access particle's area
information by calling ParticleAnalyzer()

"
:
ResultsTable rt = new ResultsTable();
:
float[] area_size_Res =
rt.getColumn(ResultsTable.AREA_FRACTION);
:
"

Also key "AREA" is not working also. I found these two
key words are defined in ResultTable.java.

However, I can access particle's X, Y coordinates
successfully.

I wonder if I need to do particular initialization to
get this area information?

thanks your time,

Michael,

error message
"
java.lang.NullPointerException
        at MTrack3_.track(MTrack3_.java:166)
        at MTrack3_.run(MTrack3_.java:116)
        at ij.IJ.runFilterPlugIn(IJ.java:196)
        at ij.IJ.runUserPlugIn(IJ.java:264)
        at ij.IJ.runPlugIn(IJ.java:116)
        at ij.Executer.runCommand(Executer.java:95)
        at ij.Executer.run(Executer.java:49)
        at java.lang.Thread.run(Thread.java:595)
"





__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

[kapilsaxena]

did you find the solution of your problem