How to separate the overlapping particles ?

[@mi]

Hi,

I am working on project in which I have to detect each cell into the sample and have to make bounding box onto the detected cells.

but I am not able to separate the joint cells and the VI is detecting three or four joint cells as a one cell and makes bounding box onto them, so I need help into separate the joint cells.

I have attached the VI and result image for better understanding.

[PeterFoerster]

I don't know your requirements for accuracy of this (+ I can't open your VI as I have 2015 installed), but I would try a mix of:

find circle: Are all your cells vaguely ciruclar? Then this function is your friend

erode/dilate: some cells look like they are barely visible. With erode and dilate you can shrink/grow the particles and potentially close gaps in the partially visible particles.

Thresholding: You get the images as gray images, right? I assume that for those circles that are partially visible, the grayscale value is different to the full ones. You could try to vary your thresholding parameters and maybe get better particles.

If you have any follow up questions, please upload an original image if possible.

[Bob_Schor]

See my reply to your original post.

Bob Schor

[@mi]

Hi,

I have attached the picture of the Vi  so you can see flow of the design.

I do not need very much accurate result but I just need to do as better as possible,

yes my cells has uncertainty in shape and roughly circular in shape.