First of all, i dont know if the following request is allowed within the guidelines of stackExchange, but i will post it and delete it if deemed necessary.

I am a student (computer science) who was tasked to build a program to count cells in an image. pretty straightforward. This task was given to me because imageJ was not accurate enough and because the researchers wanted to determine the ratio of cells that are in fibers and out of the fibers. you can see an example image normally that illustrates this (down below, not the same image as the researchers were working on, but from the internet, but analogous) .

Now, i was wondering if there's a great need for such a thing to make a website version of this program. It outputs an excel with the count of the cells and the ratio between the cells in and out of the fiber.

So , i wanted to ask people who are in this domain, if they would like to have such a program on the web. It uses deepcell as the model algrotihm. (the original website of deepcell isnt that great and doesnt distinguish between the cells in fiber and out of the fibers. )

if people do have a request for me to build something which imageJ lacks or something unrelated, i am glad to hear you out as well.

enter image description here

Note: the same question was asked in one of reddit's communities. https://www.reddit.com/r/labrats/comments/14qlza4/building_a_better_cell_detection_interface/

  • $\begingroup$ I think a service like that would be generally valuable. The problem I see is the variability that such a service would have to contend with: The cross product of the number of microscopes (and their settings), cell lines, culture conditions, and stainings is pretty big, meaning the service would get images that are very different on the pixel level. Solving that task for one particular point in that space isn't so hard. Solving it for the entire space may be. $\endgroup$ Jan 22 at 8:28


You must log in to answer this question.

Browse other questions tagged .