Re: Threshold Segmentation in TrakEM2

Posted by Albert Cardona-2 on Sep 04, 2013; 12:35pm
URL: http://imagej.273.s1.nabble.com/Threshold-Segmentation-in-TrakEM2-tp5004623p5004647.html

On Sep 4, 2013, at 8:49 AM, Stephan Saalfeld <[hidden email]> wrote:

> Hi Albert,
>
> doesn't such import script for importing foreign annotations exist
> somewhere?  I remember vaguely that similar problems were solved for
> Amira and Reconstruct:
>
> https://groups.google.com/forum/#!topic/fiji-devel/iW35mArUV4M
>
> http://fiji.sc/Reconstruct_Reader
>
> And you once created something for Raju Tomer's data preserving label
> names.  No deeper insight into the matter though---sorry.
>
> Best,
> Stephan

The importer works on labeled areas, that is, areas of homogeneous intensity value in single-channel images. As far as I understood, the images in question where originals with some areas painted on top. A custom approach is required.

Albert

>
>
> On Wed, 2013-09-04 at 01:21 +0200, Albert Cardona wrote:
>> Option 1: Create a script that processes each image so that the color is picked and kept and the rest is set to black. Set the script to every image from the right-click menu.
>>
>> Option 2: filter your images so that the color of interest pops up. Rigth-click, then Adjust Images - Adjust image filters...
>>
>> Then, use the pencil tool having an AreaList selected. See the manual.
>>
>> What I would do: write a script to pick up the colored areas and add them to an arealist.
>>
>> Albert
>>
>> On Sep 3, 2013, at 12:35 AM, arjundayal <[hidden email]> wrote:
>>
>>> Hi,
>>>
>>> I am using photoshop to highlight a certain region in my EM images and then
>>> registering these serial images using TrakEM2. I am now trying to create a
>>> 3D model of this region and am stuck on the manual segmentation step. The
>>> region that I want to segment is already segmented as it is a different
>>> color than the neighboring area. I would like to avoid having to redraw an
>>> outline around this area and was hoping for a way of telling TrakEM2 to use
>>> the obvious color difference as a way to segment the image. I have attached
>>> in example image to the post. I am interested in modeling the yellow region
>>> of the image without having to redraw all the outlines.
>>> <http://imagej.1557.x6.nabble.com/file/n5004623/sample.jpg>
>>>
>>>
>>>
>>> --
>>> View this message in context: http://imagej.1557.x6.nabble.com/Threshold-Segmentation-in-TrakEM2-tp5004623.html
>>> Sent from the ImageJ mailing list archive at Nabble.com.
>>>
>>> --
>>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>>
>> --
>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html