Thresholding segmentation and measuring shapes in ImageJ
ฝัง
- เผยแพร่เมื่อ 27 พ.ค. 2024
- The first of a series of ImageJ tutorials (an introduction to image analysis and using ImageJ). This first video covers the digital image and making measurements of object shape using thresholding and segmentation.
highly informative Craig. big love mate :)
Great video Craig!
Thanks Jonas 👍
Hi Craig, many thanks for this helpful video! I have a question related to the threshold for a series of images: I took several fluorescence images for a sample along with different stretching levels and these pictures are taken with the same conditions. I tried to determine the mean fluorescence intensity for the whole sample in each image by selecting several ROIs without defects and measuring the mean gray value within the threshold. Even though I selected one threshold scheme for these pictures, the auto thresholds for each image are different. Should I also set the same threshold for each image? Many thanks.
Hi, If the images are collected under the same conditions I would be inclined to use the same threshold values -unless there is good reason not to. Most important is that you openly report what you did and the justification. There is no right answer to this question just an observation of best practice. C
Thanks for the great explanation. Could you please make a video on intensity measurement in the stack of images?
Hi Azar, Thanks for the comment. Good idea on the stack measurements. I will add that to my list. Craig.
@@CraigDaly That would be fantastic! most of the videos are in a single image and very simple images with a couple of cells or objects. But most of the images from the confocal microscope are stacks of images from tissue including cells in the complex matrix. Thanks again.
Hello Craig thanks for the great video. If I wanted to calculate the intensity of the spots which are only located within the masks how would I go about doing this?
Hi Lia, if you use the AND function to make a new image which is = (original AND mask) using the Process/Math/AND function the resulting image will have only the data under the mask and you can then measure that. have a look at the video called 'measuring noisy images using binary masks' on the channel (th-cam.com/video/_QAYhO3Mlb4/w-d-xo.html). Good luck. Craig.
Thank you very much sir, that helped me a lot.
Hi Craig! Thanks for your clear instructions in all your videos.
I am measuring intracellular transcription factors stained using IF and wonder if there is a way of combining multiple ROIs using the image. At present, I have to manually selecting ROIs from the ROI manager and then use the OR (Combine) function, but it's hard to do because I have several hundred ROIs and some of them are so small that I can't see what I am selecting to add. Is there a way of selecting multiple ROIs and combining them by clicking on the picture itself? It seems simple enough to do, only I can't find a video or manual that shows how to do it. Thank you!
Hi, If you try using the 'add to manager' checkbox in the 'analyse particles' window, that would add all regions to the region manager. Does that get you one stage further on?
Hi Craig, thanks again for such a helpful video! I just wondered if there is a way of thresholding the image (in 8-bit) and then converting it back to RGB, so that it can be viewed in colour?
Hi Deborah. Yes, you could take your image back to RGB using the Image/Type/RGB colour option. Or can you just duplicate the original RGB image and use one copy for the 8-bit thresholding? Maybe I'm not quite understanding what you are aiming for.
@@CraigDaly I was just hoping to get rid of the background pixels using the thresholding method in your video. But to do that I need to first convert my RGB image to 8-bit. Once I've thresholded my image, I've tried to go back to the RGB colour option to restore it to red or green (in my case) but it does not go back. My aim is to get rid of the background so that I could take screenshots of it, in colour, for my project and then later on do a merge and stastical analysis. But that's okay, I may just have to do an illustrative version that I do not necessarily use for my merge/statistical analysis but more just for nice and clear diagrams in my write-up.
@@deborahgold1400 Ah ok. You should split your RGB image into the red, green and blue channels (image/colour/split channels). That gives you 3x 8-bit images. Do the background subtraction on each. Then merge them using (Image/colour/merge channels). C.
This man is intelligent in his talking style
This is a very nice video. I am trying to take the thresholded image and add each individual nucleus (or cell) to the ROI manager. Can you please recommend a strategy to achieve this? I tried edit>selection>create selection, but it makes one ROI with all nuclei, rather than a separate ROI for each nucleus. Thanks.
Hi, sorry for the delay in responding. It's an easy fix. After you have completed the segmentation and have a binary image, you would use the 'analyse particles' to extract the desired objects of a particular size/shape. If you also select 'add to manager' in the dialogue box it will add these to the ROI manager. Maybe i should do a quick video on that!
@@CraigDaly Hi Craig, Oh, that is an easy fix! This is much faster than manually drawing ROIs which is how my lab has always done microscopy analysis. Thank you very much!
Hi, how can I set on the final results the area measurement to be given in micrometers??
Hi, if you set a calibration for your image (Image/Properties) then all subsequent measurements will be in the desired units.
May I know why the mix and max threshold do not automatically be the mix and max gray value in your video?
Hi, sorry, I’d need to watch the video again. I don’t quite understand what you are asking. In some cases the minimum value might be just above 0 but that will be background signal. An automatic threshold algorithm will assume a certain amount of background - maybe values below 50 and so minimum threshold will be greater than the minimum intensity. Does that answer you question? Craig.
@@CraigDaly Hi, nice to hear from you! Since I am measuring the Gray value on H-DAB stained (Immunohistochemistry) slides. I have used the 'Adjust-threshold' function and move the slider until all the positive nucleus are selected and then the minimum threshold and maximum threshold are 0 and around 170-190 respectively in all of my images. I found that the maximum threshold value is always the same as the maximum gray value, and I am wondering if this is normal.
And I have ticked 'Limited to threshold' when setting measurement.
@@everysalalala5721 Ah, these DAB images, can I assume you are measuring dark objects on a light background? It may be that your dark objects are saturating at the 190 value. can I suggest you 'invert' the image and try thresholding but check the 'dark background' box on the threshold window. Just a guess without seeing your images. C.
Hi Craig, do you know how to calculate the collagen fragmentation index for confocal micrsoccopy?
Hi, sorry, I would need more detail about the fragmentation index. It’s not a measurement I have made so not sure. Sorry.
@@CraigDaly Can I please have your email so I can email you more detail along with pictures to explain what I mean. Thank you
@@shazlir5415 Hi, i wouldnt leave my email address here as it will get picked up by SPAM bots. Search for me at University of Glasgow. I'm easy to find. C.
@@CraigDaly ok. will do that. I'll email you soon
@@CraigDaly Thank you for replying. The fragmentation index is calculated by finding the mean area divided by the mean object number. This gives us a mean fragmentation index. I can provide you sample images of collagen if you think you can help me calculate it. It would be a huge help for my research paper.