search

wiggins-lab / SuperSegger

A completely automated MATLAB-based trainable image cell segmentation, fluorescence quantification and analysis suite, particularly well suited for high-throughput time lapse fluorescence microscopy of in vivo bacterial cells.
GNU General Public License v3.0
32 stars 20 forks source link

edit link and segmentation #2

Closed mxardre closed 6 years ago

mxardre commented 7 years ago

Hi, I am playing with supperSegger and most of the time it work like a charm. However sometimes there are some errors in the linkage or the segmentation. Would it be possible to edit the links in the supperseggerViewer? All the best Maxime And thanks again for this nice tool!

stellastyl commented 7 years ago

Hi Maxime,

we don't have a way to edit the links unfortunatelly - it's in the todo list to create a tool to edit the links, but I am not sure when we will get to it. Theoretically if you fix the wrong segments manually and re-link it should get the links right. If you notice a specific way that the linking fails, I could look a bit into our linking code and see if I can fix it.

Best, Stella

mxardre commented 7 years ago

Unfortunatly I think it is not a segmentation problem because bacteria are well segmented. If i understood how things work, I think the errors come from the fact that the bacteria are moving too much from on frame to another. I would much appreciate your idea on this, and if I can help I would be very glad. Shall I send you some sample? All the best Maxime

stellastyl commented 7 years ago

Oh okay, it will try to link them the best it can if they are moving too much, but it may not always get it right. Is your frame rate quite high? An experimental way to fix this is to take more phase images if possible, but you can keep the current frame rate for fluorescence images. I would be happy to look at the images. You can send me a couple of the images that have issues, and the _err files for those images so that I can see what the linking looks like.

mxardre commented 7 years ago

Hi Stella,

Here is a sample of what I have done :

https://dl.espci.fr/ticket/84306120e7e1e9947a259872799610e9 https://dl.espci.fr/ticket/1d72c0d67b586a8e6039cd30f532a917 https://dl.espci.fr/ticket/bf184d79f6072da1b155f77fd896936d

You can get back the file by doing in a terminal : cat testSegger.tar.gz.aa testSegger.tar.gz.ab testSegger.tar.gz.ac | tar xvf -

And I used the homemade training set attached.

As you can see there is some error but not much on the segmentation. Maybe a should retrain my settings?

Maxime

-- Dr Maxime Ardré

Laboratoire génétique de l'évolution. École supérieure de physique et de chimie industrielles de la ville de Paris 10 Rue Vauquelin, 75005 Paris

Tel : 01 40 79 44 42

Le 15/02/2017 à 23:55, Stella a écrit :

Oh okay, it is will try to link them the best they can if they are moving too much, but it may not always get it right. Is your frame rate quite high? An experimental way to fix this is to take more phase images if possible, and keep the current frame rate for fluorescence images if that's a requirement. I would be happy to look at the images. You can send me a couple of the images that have issues, and the _err files for those images so that I can see what the linking looks like.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/wiggins-lab/SuperSegger/issues/2#issuecomment-280167799, or mute the thread https://github.com/notifications/unsubscribe-auth/AMMST6H6OVoXMwgo00sq4ASMnOgChcSQks5rc4JWgaJpZM4MBtX0.

stellastyl commented 7 years ago

Hi Maxime, downloading these and unpacking them is quite slow. Could you just send 4-5 frame files (the _err.mat files that are in the /seg folder)? Just choose the frame numbers where the problem is visible. Thanks!

mxardre commented 7 years ago

Hi Stella,

Sorry for the late answer. I had to put this project in standby a bit, but now am in again. I am currently analysing datas. I realized that a was maybe expecting to much from the software. Few errors will produce outlayer that I can handle after while. Because I do not need the lineage of the colonies it should be all right.

I have made a home made training set. It seems to work more or less so I am happy with that.

Thank you very much for your help.

Keep in touch

Maxime

-- Dr Maxime Ardré

Laboratoire génétique de l'évolution. École supérieure de physique et de chimie industrielles de la ville de Paris 10 Rue Vauquelin, 75005 Paris

Tel : 01 40 79 44 42

Le 22/02/2017 à 01:48, Stella a écrit :

Hi Maxime, downloading these and unpacking them is quite slow. Could you just send 4-5 frame files (the _err.mat files that are in the /seg folder)? Just choose the frame numbers where the problem is visible. Thanks!

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/wiggins-lab/SuperSegger/issues/2#issuecomment-281531569, or mute the thread https://github.com/notifications/unsubscribe-auth/AMMST8nda4V5vr1x-CFcZmM4IhslRYxJks5re4XngaJpZM4MBtX0.

stellastyl commented 7 years ago

Okay, glad it is mostly working for your analysis. Let me know if you find any other issues. I would still like to add the manual linking at some point in the future. I will let you know when we have added that. Best, Stella

stellastyl commented 6 years ago

Hi, Just letting you know that I have added a beta-version of allowing to edit the links of cells in SuperSeggerViewerGui. If you try it out, let me know if there are any problems or if you have any comments. Best, Stella