array indices

[soubsn]

Hello,

Thank you for the great software, but I am having trouble segmenting images. My goal is to get individual cell bounding boxes within an image to then use that information as a mask to block out signal outside of cells in a separate fluorescent time-lapse. We are doing single molecule experiments. Using the masks I could get a population average for diffusion or individual cell diffusion and compare both. Currently, I can't even segment my images. It is worth noting that I'm taking Z-stacks in BF and converting those to a phase image, then analyzing that image. So there may be a problem with how this phase image looks to your software. Here is a sample image I'm trying to analyze and the error message

n.

Sample Image.zip

[pawiggins]

Hi n:

Hmmm… I didn’t get the same error you did but I did run into a problem.

I suggest you set:

CONST.superSeggerOpti.remove_debris = false;

More explicit instructions:

I would run things using processExp, rather than the gui. Make your own copy of processExp and then add the following edit:

After the CONST structure is generated, modify it by setting:

CONST.superSeggerOpti.remove_debris = false;

before BatchSuperSeggerOpti is called. Then use you new version of processExp to segment the data. Better yet, just use superSeggerOpti to directly generate the cell masks if you are doing single molecule. That is what we do in this context. You only need all the other parts of the code if you need to do time-lapse cell segmentation.

—PAW

On Nov 15, 2018, at 8:21 AM, soubsn notifications@github.com wrote:

Hello,

Thank you for the great software, but I am having trouble segmenting images. My goal is to get individual cell bounding boxes within an image to then use that information as a mask to block out signal outside of cells in a separate fluorescent time-lapse. We are doing single molecule experiments. Using the masks I could get a population average for diffusion or individual cell diffusion and compare both. Currently, I can't even segment my images. It is worth noting that I'm taking Z-stacks in BF and converting those to a phase image, then analyzing that image. So there may be a problem with how this phase image looks to your software. Here is a sample image I'm trying to analyze and the error message

n.

Sample Image.zip

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[soubsn]

Hey PAW,

I tried what you had mentioned but still got the same error. The error seems to be in the scoreNeuralNet function. I've attached the x variable used in this function.

By using the superSeggerOpti do you mean using the function itself, changing the BatchSupperSeggerOpti to superSeggerOpti in my version of processExp or using the settings in the CONST structure? I'm guessing maybe part of the problem is i'm only segmenting a single image instead of a timecourse of BF.

n.

[soubsn]

Hey PAW,

I tried the sample images you provide on your website and I get errors, with one image or many images. I'm thinking maybe there is a problem with my software

n.

[soubsn]

Hey PAW,

Problem Fixed, thank you for your help.

n.

[pawiggins]

What was the issue? So I could either fix the code or help others not have the same problem.

On Nov 19, 2018, at 2:25 PM, soubsn notifications@github.com wrote:

Hey PAW,

Problem Fixed, thank you for your help.

n.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/wiggins-lab/SuperSegger/issues/24#issuecomment-440064004, or mute the thread https://github.com/notifications/unsubscribe-auth/Amb9obm9QBT2_-2xMz6jWliXL7TQoCzUks5uwy_GgaJpZM4YgPp1.

[soubsn]

I updated to Matlab R2018b a few months ago and forgot that I couldn't find the Neural network toolbox. However, after remembering that today, I googled and found out they replaced the neural networking toolbox with the deep learning toolbox. Once I installed that it worked.

Thanks for your help

[pawiggins]

We do check for toolboxes and throw some sort of error message in checkToolboxes.m… did this fail? Or maybe you didn’t see the little error warning?

It should say: ‘Please install the neural network toolbox’ or something.

—PAW

On Nov 19, 2018, at 2:32 PM, soubsn notifications@github.com wrote:

I updated to Matlab R2018b a few months ago and forgot that I couldn't find the Neural network toolbox. However, after remembering that today, I googled and found out they replaced the neural networking toolbox with the deep learning toolbox. Once I installed that it worked.

Thanks for your help

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/wiggins-lab/SuperSegger/issues/24#issuecomment-440066172, or mute the thread https://github.com/notifications/unsubscribe-auth/Amb9oRibdCuG0GxYVtYg8Q4OvWERLsK6ks5uwzGTgaJpZM4YgPp1.

[soubsn]

There was never an warning. I do have another version of matlab on my computer maybe that is why? However once I installed the deep learning toolbox, I got the warning message saying I don't have the neural network toolbox but the program still works