Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Richard,
Metamorph may be able to do this as well. It offers two types of
tracking protocols, one called track objects and one called track points.
Using Track Objects you draw a square around a spot, define its width
and height, how far the object can travel then tell it to track. When
it looses a spot it will ask you what you want to do, whether find it
manually etc... It is a slightly automated process that requires a lot
of interaction to work correctly, and with some practice may be what you
are looking for.
A much less automated option is the Track Points routine. I think this
is supposed to be used for a different type of purpose altogether but it
might be useful. Each frame you click on the spot you want to track.
This way you are rellying only on the human eye to locate the objects.
This option probably is only good for motion measurements, not
intensity measurements, since you are not actually selecting objects but
just points.
I don't know if you have already tried these options but I figured I
would put them out there.
Steven Luke
Richard M. Parton wrote:
> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
>> ... is there any software out there that would let the
>> user simply "draw" a box round a particle / particles of interest and
>
>> then track (with centroid analysis) through subsequent images of a
>
>> time-course (The Epix - xcap software has something a bit like this
>
>> as far as I can make out but I'm not sure)?
>
>>
>> We can collect images such that neighboring particles would always be
>
>> further away than the same particle can have moved in a subsequent
>
>> images (3 to 5 fps) so the tracking algorithm could simply find the
>
>> nearest bright particle. We have 16 bit images (actually 12 bit data).
>
>>
>> I would, of course, be delighted to be proved wrong if someone knows
>
>> of the perfect automated tracking software.
>
>>
>> Thanks,
>
>>
>> Richard Parton
>
>>
>> WELLCOME TRUST CENTRE for CELL BIOLOGY
>
>> ICMB, King's Buildings, The University of Edinburgh
>
>> Mayfield Road, Edinburgh EH9 3JR, Scotland
>
>
> _____________________________________
>
>
>> Search the CONFOCAL archive at
>> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>>
>> Hello,
>>
>> If you use 16 bit digitalisation which is an option on the BIORAD it can
>> happen that you don't see much but you can work with different looking
>> window and perform all segmentation on the 16 bit images whirout a
>> compromise.
>>
>> Bye
>>
>>
>> ----- Original Message -----
>> From: "Wes Wallace" <[log in to unmask]>
>> To: <[log in to unmask]>
>> Sent: Monday, October 28, 2002 7:06 PM
>> Subject: morphometry of fluorescent images
>>
>>
>> > Search the CONFOCAL archive at
>> > http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>> >
>> > Hello Confocalists,
>> >
>> > I have a very thorny question about morphometry of fluorescent images.
>> >
>> > My question concerns the problem of contrast adjustment and
>> thresholding.
>> >
>> > In my experience, the variability of specimen brightness from one
>> > preparation to another precludes the possibility of setting fixed
>> > acquisition parameters on the confocal microscope. For example, I
>> cannot
>> > acquire all images at the same detector sensitivity, because some
>> > preparations are brighter than others, and if I did use the same
>> > detector sensitivity on all, then in some cases the objects being
>> imaged
>> > would be barely detectable, while in other cases they would be clearly
>> > detectable.
>> >
>> > This leads me to believe that the contrast of each image must be
>> > normalized after acquisition, so that the objects in each image will be
>> > represented by a similar range of pixel intensities, and a
>> threshold can
>> > be applied that is not subjectively chosen for each image.
>> >
>> > But is there a way to normalize the contrast of each image that is
>> > systematic and not biased?
>> >
>> > Sincerely,
>> >
>> >
>> > Wes Wallace
>
>> > Department of Neuroscience
>
>> > Brown University
>
>> > Providence, RI 02912 (USA)
>
>> > [log in to unmask]
>
>> >
>
>
>
>--
>
>
--
Steven J. Luke
Kimmel Cancer Center Bioimaging Facility
Thomas Jefferson University
Philadelphia, Pennsylvania
|
