LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Image processing of Microscopic image

Solved!
Go to solution
Hi sasi,

-Since,the author didn't mention perfect shape of cell, i thought the circle shape detection doesn't work. Good that you know how yeast cell look like.
-When it comes to shape detection,I didn't understand how you detected them because i am unable to do that.

I've few questions regarding this if you are ok please answer
-Since, the author gave color image, which plane did you extract?
-With what Edge threshold you are extracting circle?? In your screenshot it's 75 and with same parameters i couldn't extract the same. When i kept 50 i was ablt to detect.
-Don't you think in this method, the radius of circle needs to be very much precise otherwise it will result lot of other circles??Even the white region as circles also.

-Coming to separating touching cells in binary image, in the image author provided, there are no touching cells.
-And each cell is surrounded by white layer which will be pretty much separating (I guess) isn't it?
Thanks
uday
0 Kudos
Message 11 of 17
(721 Views)

@udka wrote:
-Since, the author gave color image, which plane did you extract?

I extracted red plane only. Its basically a gray scale image whatever the plane you extract you get almost same image except windows bar.


@udka wrote:
-With what Edge threshold you are extracting circle?? In your screenshot it's 75 and with same parameters i couldn't extract the same. When i kept 50 i was ablt to detect.

In addition with edge threshold you have another tab named with shape. Here you can limit your circle radius. You can also select other shapes also.


@udka wrote:
-Don't you think in this method, the radius of circle needs to be very much precise otherwise it will result lot of other circles??Even the white region as circles also.

above mentioned answer is applicable.


@udka wrote:
-Coming to separating touching cells in binary image, in the image author provided, there are no touching cells.
-And each cell is surrounded by white layer which will be pretty much separating (I guess) isn't it?

I do agree. But there is lot of possibilites touching together in real.

---
Silver_Shaper | CLD
Message 12 of 17
(712 Views)
Hi sasi,

Thanks for the answers.
-Red Color Plane extraction was the most crucial point here(which i missed and did L plane).
-Yes i do know that tab and the limits of circle radius and other shapes. and i am talking about the radius only.Don't you think it needs to be very precise?
Thanks
uday
0 Kudos
Message 13 of 17
(707 Views)

Udka,

Thanks udka, i decided to use the "fill the holes" approach and manage to obtain my center. 

 

Silver Shaper,

I agreed with you that yeast cells are mostly circular or elliptical. Regarding the cells touching each other, is there an suggestion to seperate the cell from each other? 

 

Lastly, I am now focusing on another area that uses the live feed images captured by my camera. However i realised the number of particles is constantly changing which I am not sure why (I am speculating whether it is due to the " left to right" flow way the camera is thresholding the image?) I was thinking will it be possible for me to set the program in a way whereby the particle will only be trapped if it fulfills the yeast area requirements (By getting the average size of a yeast cell)?  

 

As attached is the screenshot of my ImageProcessing.VI. 

0 Kudos
Message 14 of 17
(704 Views)
You can try filtering using IMAQ Particle filter, to filter based on fixed known criteria.
Try IMAQ separation for separating touching cells.
Thanks
uday
0 Kudos
Message 15 of 17
(696 Views)

Hi Udka,

 

Thanks for the suggestion. This is exactly what I wanted. However, i was wondering did i connected the Sub.Vis in the wrong order. I should always threshold the image before I try to filter off the unncessary disturbances right? In addition, Despite setting "Image Area" under the "Selection Values" with the Range of 2000-3000 area, small particles with area of only 1 are still spotted. Does this means the filter is not working? Thank you for the help all this while. As attached is the configuration i am using right now.  

 

Best Regards,

Scott

0 Kudos
Message 16 of 17
(678 Views)
Yes, Particle analysis works on binary image-so you should thresold grayscale image.
-You should use Area parameter to filter the paticle based on particle area and not on image area.
-Are you sure your yeast cell area lies within 2000-3000?? When i checked it was around 12k please check once again.
-From block diagram we cannot the control values and indicator you should provide us front panel image also to see what values you have kept.
Thanks
uday
0 Kudos
Message 17 of 17
(669 Views)