Skip to first unread message
Nicoco
Apr 13, 2016, 2:30:02 PM4/13/16
to Icy imaging
Hi,
I'm trying to use the spot detector on cell images. I already have a valid segmentation of the cytoplasms and I think I've succesfully imported them as a mask. However, on the edges of the cytoplasm are detected A LOT of spots. I believe I'm using the spot detector wrong. Here's a picture of my pipeline:
I'm trying to use the spot detector on cell images. I already have a valid segmentation of the cytoplasms and I think I've succesfully imported them as a mask. However, on the edges of the cytoplasm are detected A LOT of spots. I believe I'm using the spot detector wrong. Here's a picture of my pipeline:
Is this a limitation of the spot detector or am I doing something terribly wrong ?
Nicoco
Apr 22, 2016, 3:34:43 AM4/22/16
to Icy imaging
I'm still stuck with the same problem :(
No idea ? No one ?
No idea ? No one ?
Stephane
Apr 22, 2016, 6:44:52 AM4/22/16
to Icy imaging
Hi Nico,
Sorry you didn't got any replies for your message. It looks like you're trying to detect spot inside your cell right ? From the image i believe you should use scale 3 instead and lower sensitivity to 30 or something like like but it would be better it you could provide us an image sample so we can test it here :) Thanks.
- Stephane
Sorry you didn't got any replies for your message. It looks like you're trying to detect spot inside your cell right ? From the image i believe you should use scale 3 instead and lower sensitivity to 30 or something like like but it would be better it you could provide us an image sample so we can test it here :) Thanks.
- Stephane
Nicoco
Apr 22, 2016, 8:21:46 AM4/22/16
to Icy imaging
Thanks Stephane for taking the time to answer.
I want to detect spots inside the cells. The images I need to work on are simulated using simcep (http://www.cs.tut.fi/sgn/csb/simcep/) and I'm trying to compare the robustness of different spot detection techniques from different softwares.
Here is an example of a simulated image and its groundtruth: http://imgur.com/a/zMmLc
I need it to work using protocols (or a script, but I have no experience in Java...) because I need to test it on ~1000 images.
I want to detect spots inside the cells. The images I need to work on are simulated using simcep (http://www.cs.tut.fi/sgn/csb/simcep/) and I'm trying to compare the robustness of different spot detection techniques from different softwares.
Here is an example of a simulated image and its groundtruth: http://imgur.com/a/zMmLc
I need it to work using protocols (or a script, but I have no experience in Java...) because I need to test it on ~1000 images.
Stephane
Apr 25, 2016, 4:55:09 AM4/25/16
to Icy imaging
Hi Nicoco,
The important parameter is "compute WAT considering ROI" but it looks like you checked it and your protocol seems correct.
I tried to replicate it and i could obtain the expected result so i'm joining it to the message so you can see the differences with your :)
Hope that helps !
- Stephane
The important parameter is "compute WAT considering ROI" but it looks like you checked it and your protocol seems correct.
I tried to replicate it and i could obtain the expected result so i'm joining it to the message so you can see the differences with your :)
Hope that helps !
- Stephane
Nicoco
Apr 26, 2016, 5:53:06 AM4/26/16
to Icy imaging
Thanks!
My problem was that I had to link my sequence to "input sequence" and to "input sequence ROI".
My problem was that I had to link my sequence to "input sequence" and to "input sequence ROI".
Reply all
Reply to author
Forward
0 new messages