Prev: fast decoupled load flow analysis
Next: Error message with imread and imshow
From: Dipen on 13 Feb 2010 03:20
I am trying to code a program for counting the number of round stem cells from an microscopic images.

1.) The cells are of two types.. Live (white, bright) & Dead (dark blue stained).

2.) Cells have to be identified only from the 4*4 matrix. All other cells outside 4*4 matrix are to be neglected.

3.) Can you provide the info in brief what would be the appropriate steps or algorithm for counting total cells as well as differentiating the live and the dead by a specific mark on the image?

Those who are interested for developing the code with me can send there Email ID's on <dppatel_184u(a)yahoo.com> so that i can provide you the Raw Images Of Stem Cells.
From: ImageAnalyst on 13 Feb 2010 07:42
Dipen:
Extract the red channel of your images like this
redChannel = rgbImage(:,:,1);

Then use it as the monochrome input image in this demo:
http://www.mathworks.com/matlabcentral/fileexchange/25157

And upload your photos to http://drop.io if you want me to look at
them.


From: Dipen on 15 Feb 2010 02:34
Thank you so much for the reply.. I vl try it and let u knw as soon as possible.

You can upload the images from http://drop.io/dipen2255268

One more favor is needed, as am new to the Matlab, help me out from the beginning of the code in detail if you can.
From: ImageAnalyst on 15 Feb 2010 13:01
Dipen :
Are those two images identical?
Can you get the image without that grid in the image?
It's not clear what's live and what's dead. I only see one small
white speck - is that the one live cell? I see numerous (hundreds of)
dark specks of all different sizes and shapes and contrasts - are they
ALL dead cells that need to be counted. This could be tough because
some of them are so small and close to the background intensity that
they blend in with the background. You might try a tophat or
bottomhat filter on the image to find them.
From: Dipen on 16 Feb 2010 02:55
Yes Sir, those two images are identical.

I have uploaded the images at http://drop.io/dipen2255268 just for your knowledge.. check it out..

In the images there are many live cells which u will see if u ZOOM into the image.
and there are dead cells which are scattered in the image and also there are 2 clumps of dead cells on center left.

And those lines are the partitions of Neubauer Chamber slide(check out the link which shows the image of actual neubauer counting chamber). There are main 4 blocks. Namely 1, 2, 4 & 5 blocks.
1 block = 4X4 matrix of 16 boxes.

1.) The main aim is to count the cells only in 16 boxes (4 X 4 matrix) which makes 1 out of 4 blocks.

Each live cell has a definite diameter..say for example 10 microns.
We have to provide a range of 8 to 10 microns and all those white specks which fall in this range and are inside the block (4 x 4 matrix) are to be considered and others are to be neglected in counting and differentiating.

So, total number of cells from the four blocks are to be counted and then their Average is considered to be the final count. ( Easiest way to count is that we crop the 4 boxes from a single block at a time instead of cropping all 16 boxes individually and then count live as well as dead and then the average of those 4 boxes. That's it.)

2.) To differentiate the live cells from the dead and give their respective count.
-live cells are transparent white with a definite boundary.
-dead cells are stained blue dots.

ANY Queries let me know m ready to help at every step. Am trying my best to solve it coz its important for me to complete it on time within 4 months.
 |  Next  |  Last
Pages: 1 2
Prev: fast decoupled load flow analysis
Next: Error message with imread and imshow