1

Before start, please note that this question is not the duplicate of this one. In cells, some of the proteins are intentionally tagged with fluorescence to see their shine under a special kind of microscope. My part is to find this shine in images. There are two types of images, one of them is bright field and the other one is so called fluorescence image which includes valuable information.

I tried many techniques to find a shine on cell; however, I cannot find anything except some noise. Note that, if there is a shine, it will be really small; however, it should be the brightest thing in image. I find this task really challenging and want to share with you. Again note that, this cell definitely includes fluorescence, there is no doubt about it.

I'm asking two things here. One is how can I find a shine in these images. The second is how do you know that it's not noise. The images are presented below. I put the bright field images just to give you an idea about the cell.

enter image description here enter image description here

enter image description here enter image description here

So far, I tried:

  • Color space changing
  • Contrast enhancement
  • Maxima analysis
  • Threshold

Bigger images:

First Fluorescence Image:

enter image description here

Expected location of Fluorescence in First image(marked with spray):

enter image description here

Second Fluorescence Image:

enter image description here

Expected location of Fluorescence in Second image(marked with spray):

enter image description here

Third Fluorescence Image:

enter image description here

Expected location of Fluorescence in Third image(marked with spray):

enter image description here

Community
  • 1
  • 1
cagatayodabasi
  • 762
  • 11
  • 34
  • Can you please clarify what each of the four images is and also provide additional marked up images with a circle around the shine areas you are looking for? – Mark Setchell Oct 07 '16 at 21:47
  • The first column is bright field images and there is no information in them. I put them just to show you the cell. The second column is taken under special filters and the shine should be in there. I updated the question explanation and added new bigger images. – cagatayodabasi Oct 07 '16 at 22:51
  • Your images don't seem to agree with your words. You say that if the shine is there, it will be the brightest pixel - yet none of the images have the brightest pixel where you have marked the shine should be? You also say you have tried Maxima Analysis, but yet again, if your shine pixels are the brightest ones, how did Maxima Analysis fail? – Mark Setchell Oct 08 '16 at 00:00
  • In ideal case, they should be the brightest. However, these images are taken under special type of filters(I don't know the details, because I'm not a biologist but am an image processing guy.). The images seems dark, but I analyzed some of them and if I changed the color space from RGB to Lab, I can see something like the walls of chips(because they can reflect the light coming from microsocope); however they are not cells. Note that, I asked the question, since I guess the noise level(actually SNR) is way too high. So, maybe someone from community can find something among this noise. – cagatayodabasi Oct 08 '16 at 09:20

0 Answers0