Enough for cell counting?

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

rgcalsaverini

Member
Joined
Aug 31, 2011
Messages
19
Reaction score
0
Location
SP
I have a strange feeling that I am about to get shouted at by the microbiologists here... But I'll ask it anyway:

Would This be enough for counting yeast cells?
The thing is, I want to automate the process and use computer vision algorithms to do all counting, so basically I need decent res. pictures of yeast coming trough USB, but I know close to nothing about microscopy. I was told that around 100x would be enough, is that true? If that product is crap, what would you suggest instead?
 
Last edited by a moderator:
yeast are tiny. I would imagine you are going to want far more magnification than 100X. Do you really plan on actual counting? If so, you've got quite a task ahead of you. I'd say plan a full weekend to count a slide. They're plentiful.
 
yeast are tiny. I would imagine you are going to want far more magnification than 100X. Do you really plan on actual counting? If so, you've got quite a task ahead of you. I'd say plan a full weekend to count a slide. They're plentiful.

Really? That tiny?
Well, actually, I'm not going to count it, I'll make a computer vision algorithm that will do all the counting automagically within minutes :)

In case you are curious:
http://en.wikipedia.org/wiki/Computer_vision
 
You will need more like 200x, as 100x is a bit to small.

You can count cells under your setup, but you do not knwo volume. Cell counting is usually accomplished with a hemocytometer. That gives a known volume to count in. Absolute standard microbiological yeast equipment.
 
Yeah, but you have to figure out how to get the program to count what is inside the hymocytometer grid and nothing outside of it.

This might work if the grid took up 100% of the FOV of the microscope, but the only hemocytometers grids I've seen are square and microscope FOV is round.

If ImageJ has a limit feature where you can draw a box around the hemo grid, you might be in business....
 
You could do it with a micropipette. Place on slide and count. Then figure how many drops make 1mL and total volume. Would give a rather accurate estimate that is probably accurate to 10%.
 
Yeah, but you have to figure out how to get the program to count what is inside the hymocytometer grid and nothing outside of it.

This might work if the grid took up 100% of the FOV of the microscope, but the only hemocytometers grids I've seen are square and microscope FOV is round.

If ImageJ has a limit feature where you can draw a box around the hemo grid, you might be in business....

Trivial. If you can detect a yeast cell, you can detect the grid, and you don't even need to do that. The grid isn't going to move, you just figure out what pixels make up the grid, and run your algorithm on those, and you actually don't even need to do that. Determine how much of the FOV the grid takes up, and calculate a multiplier. Now you can just use the entire FOV along with your multiplier.
 
For yeast counting, you really want about 400x. I'd say 100x would be a pain.

At magnification, you want a decent X/Y table to move the sample around. I don't see anything like that in the pic you posted.

There is a already a program designed to find cells in an image. Google ImageJ

Holding the cell phone over the eyepiece is a royal pain. Better to take a cheap webcam, remove the lens (must do this), remove the eyepiece lens from your microscope, and shove the webcam into the tube. Works, I've done this for a long time. Better yet, get an adapter for a DSLR camera and just mount your camera to the microscope eyepiece tube.

There are excellent time-lapse programs that will take advantage of any webcam and allow you to make a video over long time intervals (basically creating an intervalometer).
 
You can hold a cell phone over the eye piece of a microscope, upload to your computer and use this to count it:
http://rsbweb.nih.gov/ij/

For yeast counting, you really want about 400x. I'd say 100x would be a pain.

At magnification, you want a decent X/Y table to move the sample around. I don't see anything like that in the pic you posted.

There is a already a program designed to find cells in an image. Google ImageJ

Holding the cell phone over the eyepiece is a royal pain. Better to take a cheap webcam, remove the lens (must do this), remove the eyepiece lens from your microscope, and shove the webcam into the tube. Works, I've done this for a long time. Better yet, get an adapter for a DSLR camera and just mount your camera to the microscope eyepiece tube.

There are excellent time-lapse programs that will take advantage of any webcam and allow you to make a video over long time intervals (basically creating an intervalometer).

I'm not sure that ImageJ is a good beginner's program to use... It shouldn't be too difficult to count the cells by "hand". We do this in our labs all the time. Getting the "real" equipment will be pricey since a Neubauer hemacytometer is a precision instrument. A microscope should be affordable, I suppose (eBay?). For cell counting we never use the built-in table to shift the hemacytometer. Maybe there are less "professional" alternatives out there since often you pay premium just because the item has "professional" or "industrial" attached to it.
 
I'm not sure that ImageJ is a good beginner's program to use... It shouldn't be too difficult to count the cells by "hand". We do this in our labs all the time. Getting the "real" equipment will be pricey since a Neubauer hemacytometer is a precision instrument. A microscope should be affordable, I suppose (eBay?). For cell counting we never use the built-in table to shift the hemacytometer. Maybe there are less "professional" alternatives out there since often you pay premium just because the item has "professional" or "industrial" attached to it.

I agree, I count manually as well. Easy if you can dilute properly. I mentioned the software only because the OP brought it up.

You can get a Neubauer hemacytometer on ebay for about $30. I did.

I use the X/Y stage a lot! I'm surprised you can do it without that, but I do believe you. Amscope makes what looks like decent scopes for low dollars. I bought a used Olympus on ebay; wasn't cheap.
 
I agree, I count manually as well. Easy if you can dilute properly. I mentioned the software only because the OP brought it up.

You can get a Neubauer hemacytometer on ebay for about $30. I did.

I use the X/Y stage a lot! I'm surprised you can do it without that, but I do believe you. Amscope makes what looks like decent scopes for low dollars. I bought a used Olympus on ebay; wasn't cheap.

Cool! I am used to paying $200+ for a new Neubauer :mad:

I guess I don't drink a lot of coffee so I have steady enough hands to move the chamber. Are they that much smaller than the average mammalian cell? Because I always focus in on one set of 4x4 squares, count, and move on to the next...
 
I bought a used scope and a hemacytometer on eBay for just over $100. We will see how it goes.
 
Back
Top