Hi Craig! Thank you so much for this great video. I was able to learn a lot from it. Not sure if you will see this comment of mine but thought it would worth the try. I have to count neutrophil cells in zebrafish larval's hindbrain using ImageJ software. struggle at the moment as the neutrophil cells do not have clear boundaries e.g like colonies but appear in patches instead. I just wonder if I can get some help which I greatly appreciate :)
Hi, email me a sample image and I will try and suggest something. I don’t want to leave my email address here for the spam BOTs to pick up. Search for me at University of Glasgow. I’m easy to find. C.
Thanks for all your videos Craig, really great. Any tips on following a similar process on 32-bit float images? Perfoming an image calculation of the 32-bit original AND 8-bit mask leads to odd results (not identical to converting the original to 8-bit). Would like to keep my raw 32-bit scale rather than converting to a 0-255 integer scale
Actually, I've used a workaround - dividing the 8-bit mask by 255 so I end up with a mask of 1s and 0s - then multiplying the original 32-bit by that mask - which seems to solve my query. Thanks again
Hi Christos, that’s good you found a workaround. I only have access to ImageJ on an iPad at the moment which is not ideal. I did manage to do an AND using an 8-bit mask and 32-bit image which seemed to work ok. I was going to ask for a bit more info but you solved it. Great. Thanks for watching. Craig.
Thanks for the video! Does the process of binary masking (changing pixel values throughout this process) still allow for measuring intensity, for example, since the end result seems to have matching pixel values, or is measuring only suitable for the original raw image?
Hi, yes if you are measuring intensity you would need to use the original image when you do the binary AND process. However, you could use a lowpass/enhanced image to make the binary mask but then apply that to the original image. Hope that makes sense.
Hi Paul, thanks for the comment. I’m glad it helped. Yes, when you apply a threshold to a z-series ImageJ will ask if you want to apply it to all images (or rather, you check that box). Please share links to the channel if you know anyone that might be interested.
Thanks again for making these great videos. Can we use the dilated Binary image directly to analyze particles? actually, I did not get what is the purpose of the last step (image calculator)?
Hi, yes you could measure the dilated binary image using analyse particles but you must remember that the shape could be very slightly different from the original. Only very slightly though. Just make sure that you list your methods. The image calculator allows us to add and subtract images or use binary operators. In the example I used a ‘logical AND’; meaning take two images (the original and the mask) to create a new image which contains only the pixels that are in image 1 AND image 2. Another use is to show everything which is in image 1 but NOT in image 2. That’s good for seeing differences between images. Alternatively, combine images using the OR function. Try it yourself with some images that are the same size and it should make sense.
Hi, Sorry it's too long ago since I made this. Im not sure what image 3 is. If that is the raw (original) image that has been extracted using the mask (ie output of: 'BinaryMask AND OriginalImage') then yes, you can measure intensity of that result.
@@CraigDaly yes, that is what I meant. Thank you. I am not an experienced user of image J, would it be possible to apply the same method in a 1000 images without having to process them individually?
@@dumisilelumkwana2683 Hi, Yes, if all of the images are very similar you could build a macro to analyse multiple images. This just requires you to record all of the moves for one set of data and then save that as a macro. You can then apply the macro to all images in a directory.
Thank you. I need a bit different thing but this buide helped me a lot. However i have a question, can i save picture without background? What i need is outline nanograins and outline function/ find edges function/ or this method do with some error (sometimes it outline 5-10 grains as 1 grain). SO i want to save picture of edges without background and use it as second layer in, for example GIMP so i would be able to separate those grains manually
Hi, Difficult to picture the exact problem without seeing the images. However, my first thought is to use some kind of image addition or subtraction. I dont use GIMP as photoshop is my go to for image manipulations. If you do an edge detection in imageJ (sobel filter maybe) and then try an AND of the two images would that work? Or in photoshop I would just use the selection tool to pick all of the background, invert the selection to get the NOT-background and copy paste to a new image. Sorry if I have misunderstood the problem.
I need to measure the intensity for my experiments. I am a bit confused. Should i "multiply" or "add" my original image with the threshold image. can anyone please let me know?? Thanks
Hi, sorry for the delay, I’ve been away from RUclips for a while. If I understand correctly you should only AND the raw image and the mask. This reveals the data under the mask which you can then measure as average intensity or total intensity. I would avoid ADD or MULTIPLY.
Holy heck, bless you and your mother
Thanks a lot Craig ..... very important videos with valid and relevant content
That was incredible
Great video, Graig.
Thank you.
Thanks Frank.
thank you so much for this great video. please do share more .. really great stuff :)
Thanks for watching. I will keep them coming. C.
Thank you Craig! Saved my life! doing image processing on fibers and found this really helpful thanks!
Great, glad it was useful. Craig.
thank you for the video. I appreciate it and would like to see more image J videos
Thanks Murat, i will try and keep them coming.
Great Videos Craig! This information is also suer helpful for Earth Scientists! Please keep up the good work. Loving Greetings from Germany.
Thanks Bryce, it’s good to know that they are useful. Best wishes.
THANK U CRAIG
Hi Craig! Thank you so much for this great video. I was able to learn a lot from it. Not sure if you will see this comment of mine but thought it would worth the try. I have to count neutrophil cells in zebrafish larval's hindbrain using ImageJ software. struggle at the moment as the neutrophil cells do not have clear boundaries e.g like colonies but appear in patches instead. I just wonder if I can get some help which I greatly appreciate :)
Hi, email me a sample image and I will try and suggest something. I don’t want to leave my email address here for the spam BOTs to pick up. Search for me at University of Glasgow. I’m easy to find. C.
Thanks for all your videos Craig, really great. Any tips on following a similar process on 32-bit float images? Perfoming an image calculation of the 32-bit original AND 8-bit mask leads to odd results (not identical to converting the original to 8-bit). Would like to keep my raw 32-bit scale rather than converting to a 0-255 integer scale
Actually, I've used a workaround - dividing the 8-bit mask by 255 so I end up with a mask of 1s and 0s - then multiplying the original 32-bit by that mask - which seems to solve my query. Thanks again
Hi Christos, that’s good you found a workaround. I only have access to ImageJ on an iPad at the moment which is not ideal. I did manage to do an AND using an 8-bit mask and 32-bit image which seemed to work ok. I was going to ask for a bit more info but you solved it. Great. Thanks for watching. Craig.
Thanks for the video! Does the process of binary masking (changing pixel values throughout this process) still allow for measuring intensity, for example, since the end result seems to have matching pixel values, or is measuring only suitable for the original raw image?
Hi, yes if you are measuring intensity you would need to use the original image when you do the binary AND process. However, you could use a lowpass/enhanced image to make the binary mask but then apply that to the original image. Hope that makes sense.
Craig you've saved my life haha!
Is there a way to adjust threshold on a z-stack of images?
Hi Paul, thanks for the comment. I’m glad it helped. Yes, when you apply a threshold to a z-series ImageJ will ask if you want to apply it to all images (or rather, you check that box). Please share links to the channel if you know anyone that might be interested.
Thanks again for making these great videos. Can we use the dilated Binary image directly to analyze particles? actually, I did not get what is the purpose of the last step (image calculator)?
Hi, yes you could measure the dilated binary image using analyse particles but you must remember that the shape could be very slightly different from the original. Only very slightly though. Just make sure that you list your methods. The image calculator allows us to add and subtract images or use binary operators. In the example I used a ‘logical AND’; meaning take two images (the original and the mask) to create a new image which contains only the pixels that are in image 1 AND image 2. Another use is to show everything which is in image 1 but NOT in image 2. That’s good for seeing differences between images. Alternatively, combine images using the OR function. Try it yourself with some images that are the same size and it should make sense.
@@CraigDaly Thank you Craig for taking the time and answering my question. it was very helpful.
Hi Craig, thank you for this video. Just wanted to make sure if I want to calculate the intensities, I would the use image 3?
Hi, Sorry it's too long ago since I made this. Im not sure what image 3 is. If that is the raw (original) image that has been extracted using the mask (ie output of: 'BinaryMask AND OriginalImage') then yes, you can measure intensity of that result.
@@CraigDaly yes, that is what I meant. Thank you.
I am not an experienced user of image J, would it be possible to apply the same method in a 1000 images without having to process them individually?
@@dumisilelumkwana2683 Hi, Yes, if all of the images are very similar you could build a macro to analyse multiple images. This just requires you to record all of the moves for one set of data and then save that as a macro. You can then apply the macro to all images in a directory.
@@CraigDaly great, thank you so much for your help
Thank you. I need a bit different thing but this buide helped me a lot. However i have a question, can i save picture without background? What i need is outline nanograins and outline function/ find edges function/ or this method do with some error (sometimes it outline 5-10 grains as 1 grain). SO i want to save picture of edges without background and use it as second layer in, for example GIMP so i would be able to separate those grains manually
Hi, Difficult to picture the exact problem without seeing the images. However, my first thought is to use some kind of image addition or subtraction. I dont use GIMP as photoshop is my go to for image manipulations. If you do an edge detection in imageJ (sobel filter maybe) and then try an AND of the two images would that work? Or in photoshop I would just use the selection tool to pick all of the background, invert the selection to get the NOT-background and copy paste to a new image. Sorry if I have misunderstood the problem.
@@CraigDaly thank you i already solved the problem and my solution is very similar to your message. Again, thank you for reply
I need to measure the intensity for my experiments. I am a bit confused. Should i "multiply" or "add" my original image with the threshold image.
can anyone please let me know??
Thanks
Hi, sorry for the delay, I’ve been away from RUclips for a while. If I understand correctly you should only AND the raw image and the mask. This reveals the data under the mask which you can then measure as average intensity or total intensity. I would avoid ADD or MULTIPLY.
Dear Craig i have 2 images i want to change thise pictures to good quality one. Can you help me ?
Hi, sure. Email me at University of Glasgow and i will try and offer some advice. Craig.