Thanks for the great videos. I would like to measure the length of cilia after watching the video, but is there any way to apply the actual measurements? Also, when I analyzed with the first method, I didn't see the number of each object, do you know how to solve it?
Hi @dongwooseo4366! The number of objects that were thresholded can be found in the log window that appears. See 2:51 of the video. For length of cilia in z stacks, you might want to consider skeleonization. ruclips.net/video/UnYWdQVtxfc/видео.htmlsi=4q15dHlUYy1EF62R I also heard about a new plugin called CiliaQ, which measures ciliary shape, length, and fluorescence in 2D, 3D, and 4D images. I haven't tired this yet, but you might want to explore this as an option.
Hi, for the 3D manager part (the second method), can I add a binary segmented image, which was previously segmented using Labkit, Ilastik, etc. instead of relying on the segmentation of the ROI manager? That would be more accurate than using thresholding to segment my image.
Hello @acufssdz12. Yes, you can, however, you will need to first convert the binary image stack into a label image stack. You can use the 3D Segmentation button on the 3D Manager, just keep the low threshold at 128 and the high at 255. This will output a label stack, which you can then "Add" into the 3D Manager as objects.
Hi Johanna! thanks for this tutorial in particular and all other videos in your channel. Very instructive! I have a question. In 2D intensity measurement is usually applied a background correction by subtracting the background mean intensity to the whole image. Should this background subtraction also be applied to 3D measurements? I used the SNT plugin to create a 3D selection for some axons and with the steps in this video, I managed to make measurements with ROIManager3D restricted to the axons. However I'm not sure how I should proceed with a 3D background subtraction. Any advice?
Thanks @Miquel for your support. I appreciate it a lot. Background subtraction helps in the segmentation process. If you are able to successfully isolate the axons from the background (without any pre-processing), then object detection is not too much of an issue. But, if you are doing intensity measurements, just keep in mind that background intensity adds to the fluorescence signal intensity. For z stacks, you can get the average intensity of the background (in all slices), then subtract this value from your intensity measurements. Or you can also just do standard rolling ball radius background subtraction on your original 3D image (applying this to all slices) and do measurements from there.
Is there a way to set a slice thickness, so the volume doesn't treat the slices as just being stacked on top of each other with no slice thickness since this will skew the results of the volume of the cells? Also what exactly does Integrated Density mean?
Hi @Praneel Sunkavalli. If your images were acquired on a confocal microscope, slice thickness would depend on the size of the pinhole. If you have an image open, go to Image> Properties. The voxel depth determines the spacing between slices. Integrated density is the sum of the values of the pixels in the image or selection (ROI). This is equivalent to Area x Mean Gray Value of the ROI.
Hi! How can I set a size filter using the 3D ImageJ suite plugin? I have a z-stack with cells of many dimensions and need to analyze only cells in a certain range. thank you!
Hi @Sara Toros. This followup video describes how to set a size filter (among others) using the 3D Manager: FIJI (ImageJ): Using 3D Objects Counter and 3D Manager to Measure Cells in a Z Stack ruclips.net/video/dmJSRFBmVXg/видео.html Hope it helps.
@@johanna.m.dela-cruz yes it was really helpful thank you. If I can, I would ask you one more thing because I'm really struggling with detecting some cells. I managed to detect all the cells of interest with the 3D object counter, but I'm trying to do it with the suite plugin since I need the ROIs.. I set the same threshold I used with the counter (as you do in the video) but the suite plugin doesn't seem to create an ROI for all the cells/objects of interest (even if they are in the map it creates after doing '3d segmentation'). I don't understand why this happens and how to improve my analysis.. I hope I was clear enough and thank you for your help!!
@@saratoros6429 Hi Sara. Do you more or less have the same number of cells? Did you, perhaps, exclude edges on either one of the 3D objects counter or 3D Manager? An ROI on each slice is most likely a portion of the whole ROI...if you use a different LUT, you might be able to see the lower intensity voxels of your segmented stack. I'll be happy to assist you further if we communicate via email.
Hi @janne4440, you will need to add the 3D Image Suite update site in Fiji, as indicated in 0:23 of the video. If you are using ImageJ, I suggest you download Fiji instead.
@@johanna.m.dela-cruz Thanks for your swift reply, do you have a tutorial measuring distance from object to border? For instance, I have FISH DNA probe signal within nuclus, and need to know how close it is to the nuclear membrane(DAPI).
Hi Zen Borg. If your measurements have 0 values, this means that they were not included in the measurements set under Analyze >3 D OC Options. I included "Centroid" in my measurements, so the x, y and z coordinates of these centroids have measurements in the results table. The parameters you check depend really on what you want to measure, so if you don't need centroid measurements then you don't have to tick the check box.
@@johanna.m.dela-cruz thank you! I am trying to track the height of my contact line of my droplet over time. I had noticed the contact line is black compared to the rest of my background and so I ticked only centroid and Maximum grey value. From my understanding, this will then mostly plot the points at the positions whereby the regions are dark. Annoyingly, y-axis is inverted in which I had to correct for.
@@zenborg879 oh, so your image stack is not of optical sections, but a time series. Perhaps you might want to check my other video on tracking. It might help you more with your analysis goal. ruclips.net/video/kVYBiHqNmzo/видео.html
Hi Safa. A size filter can be estimated from your image stack. If there are objects that are less than the size you set (this would be in pixels), they won’t be included in the segmentation. I can discuss this more with you via email.
@@johanna.m.dela-cruz thank you! Do the different colors mean something? Are they grouped by some parameter like volume? Can we get a heatmap key if that’s the case?
@@safa93 just quoting this: “In a label image, each object is represented by different integer values. For instance, the object #1 in a label image will be made from all the pixels that have a pixel value of 1, over a black background of 0. Object #2 will have the pixel value 2, etc.”
Thanks for the great videos. I would like to measure the length of cilia after watching the video, but is there any way to apply the actual measurements? Also, when I analyzed with the first method, I didn't see the number of each object, do you know how to solve it?
Hi @dongwooseo4366! The number of objects that were thresholded can be found in the log window that appears. See 2:51 of the video. For length of cilia in z stacks, you might want to consider skeleonization. ruclips.net/video/UnYWdQVtxfc/видео.htmlsi=4q15dHlUYy1EF62R
I also heard about a new plugin called CiliaQ, which measures ciliary shape, length, and fluorescence in 2D, 3D, and 4D images. I haven't tired this yet, but you might want to explore this as an option.
@@johanna.m.dela-cruz Thank you for your answer. Have a great day.
Hi, for the 3D manager part (the second method), can I add a binary segmented image, which was previously segmented using Labkit, Ilastik, etc. instead of relying on the segmentation of the ROI manager? That would be more accurate than using thresholding to segment my image.
Hello @acufssdz12. Yes, you can, however, you will need to first convert the binary image stack into a label image stack. You can use the 3D Segmentation button on the 3D Manager, just keep the low threshold at 128 and the high at 255. This will output a label stack, which you can then "Add" into the 3D Manager as objects.
Hi Johanna! thanks for this tutorial in particular and all other videos in your channel. Very instructive!
I have a question. In 2D intensity measurement is usually applied a background correction by subtracting the background mean intensity to the whole image. Should this background subtraction also be applied to 3D measurements? I used the SNT plugin to create a 3D selection for some axons and with the steps in this video, I managed to make measurements with ROIManager3D restricted to the axons. However I'm not sure how I should proceed with a 3D background subtraction. Any advice?
Thanks @Miquel for your support. I appreciate it a lot.
Background subtraction helps in the segmentation process. If you are able to successfully isolate the axons from the background (without any pre-processing), then object detection is not too much of an issue. But, if you are doing intensity measurements, just keep in mind that background intensity adds to the fluorescence signal intensity. For z stacks, you can get the average intensity of the background (in all slices), then subtract this value from your intensity measurements. Or you can also just do standard rolling ball radius background subtraction on your original 3D image (applying this to all slices) and do measurements from there.
Is there a way to set a slice thickness, so the volume doesn't treat the slices as just being stacked on top of each other with no slice thickness since this will skew the results of the volume of the cells? Also what exactly does Integrated Density mean?
Hi @Praneel Sunkavalli. If your images were acquired on a confocal microscope, slice thickness would depend on the size of the pinhole. If you have an image open, go to Image> Properties. The voxel depth determines the spacing between slices.
Integrated density is the sum of the values of the pixels in the image or selection (ROI). This is equivalent to Area x Mean Gray Value of the ROI.
I like the video and the bg song haha
Thanks Louise. Watch and learn while listening to some music….that’s my jam. 😁
Hi! How can I set a size filter using the 3D ImageJ suite plugin? I have a z-stack with cells of many dimensions and need to analyze only cells in a certain range. thank you!
Hi @Sara Toros. This followup video describes how to set a size filter (among others) using the 3D Manager: FIJI (ImageJ): Using 3D Objects Counter and 3D Manager to Measure Cells in a Z Stack
ruclips.net/video/dmJSRFBmVXg/видео.html
Hope it helps.
@@johanna.m.dela-cruz yes it was really helpful thank you. If I can, I would ask you one more thing because I'm really struggling with detecting some cells. I managed to detect all the cells of interest with the 3D object counter, but I'm trying to do it with the suite plugin since I need the ROIs.. I set the same threshold I used with the counter (as you do in the video) but the suite plugin doesn't seem to create an ROI for all the cells/objects of interest (even if they are in the map it creates after doing '3d segmentation').
I don't understand why this happens and how to improve my analysis..
I hope I was clear enough and thank you for your help!!
@@saratoros6429 Hi Sara. Do you more or less have the same number of cells? Did you, perhaps, exclude edges on either one of the 3D objects counter or 3D Manager? An ROI on each slice is most likely a portion of the whole ROI...if you use a different LUT, you might be able to see the lower intensity voxels of your segmented stack. I'll be happy to assist you further if we communicate via email.
@@johanna.m.dela-cruz thank you so much, I sent you an email!
When I use the 3D object counter I get an error which says java.lang.ArrayIndexOutOfBoundsException. Anyone know how to fix this?
Hi @scottmoyer4024. What type of image files are you working with? How large are they? Are your objects segmented correctly?
Hi Johanna, thanks a lot, i like your videos, but i could not find 3D or 3D manager plugins, any tips_
Hi @janne4440, you will need to add the 3D Image Suite update site in Fiji, as indicated in 0:23 of the video. If you are using ImageJ, I suggest you download Fiji instead.
@@johanna.m.dela-cruz Thanks for your swift reply, do you have a tutorial measuring distance from object to border? For instance, I have FISH DNA probe signal within nuclus, and need to know how close it is to the nuclear membrane(DAPI).
@@janne4440 I believe Geodesic Distance maps might be useful: ruclips.net/video/0BTOAg5eagg/видео.html
Nice video. However, I do not get any of the x, y, z coordinates whereby they all appear 0. How do I set that and retrieve them like yours on 3:02?
Hi Zen Borg. If your measurements have 0 values, this means that they were not included in the measurements set under Analyze >3 D OC Options. I included "Centroid" in my measurements, so the x, y and z coordinates of these centroids have measurements in the results table. The parameters you check depend really on what you want to measure, so if you don't need centroid measurements then you don't have to tick the check box.
@@johanna.m.dela-cruz thank you! I am trying to track the height of my contact line of my droplet over time. I had noticed the contact line is black compared to the rest of my background and so I ticked only centroid and Maximum grey value. From my understanding, this will then mostly plot the points at the positions whereby the regions are dark. Annoyingly, y-axis is inverted in which I had to correct for.
@@zenborg879 oh, so your image stack is not of optical sections, but a time series. Perhaps you might want to check my other video on tracking. It might help you more with your analysis goal. ruclips.net/video/kVYBiHqNmzo/видео.html
Hi! How do you determine what size filter to use?
Hi Safa. A size filter can be estimated from your image stack. If there are objects that are less than the size you set (this would be in pixels), they won’t be included in the segmentation. I can discuss this more with you via email.
@@johanna.m.dela-cruz Hi Dr. Johanna! What do the different colors on the objects map depict?
@@safa93 The objects map is basically a segmented (label) image. Each color represents a different object.
@@johanna.m.dela-cruz thank you! Do the different colors mean something? Are they grouped by some parameter like volume? Can we get a heatmap key if that’s the case?
@@safa93 just quoting this: “In a label image, each object is represented by different integer values. For instance, the object #1 in a label image will be made from all the pixels that have a pixel value of 1, over a black background of 0. Object #2 will have the pixel value 2, etc.”