Відео

Hi @BCBNarjis. How big are your images? What type are they (2D, 3D, 4D)? Are you using Windows or do you have a Mac?

@johanna.m.dela-cruz the 3D image is 108mb, and the 2D image is 4 mb. I tried both 3D and 2D threshold but they both crash the app without any prompt. I am on Windows.

Hi @putusixtinadewandari4043. Do you mean your image is made up of several circular objects? Have you tried thresholding?

@johanna.m.dela-cruz not yet, how to do it? can you explain, please🙏

Hi @lokiiiii3333. Do you mean put your images together as a montage or merge all these images together? Segmentation is just a way to make object identification and analysis easier.

@@johanna.m.dela-cruz thanks for the reply! what i meant was to merge 5 diff. images into one.

@ You should first convert your RGB images to 8-bit (Image - Type - 8 bit). Then, select Image - Color - Merge. The colors can be modified via LUTs.

@@VaishnaviK-j5j Thanks for watching. In order to merge 2 different stacks, they have to have the same bit depth. You can change one of them to match with the other stack. Go to Image - Type - to do this.

@@johanna.m.dela-cruz Got it! It works now. Thanks so much:) Another question I had was related to why I'm unable to view the numbers on the 3D projection despite selecting "show numbers" on the maps parameters. Another strange thing is that when I view the results including volume, centroid volume etc I'm unable to see what serial numbers they correspond to. Sorry to bombard you with all these questions haha:')

@@VaishnaviK-j5j Did you check Statistics on the 'Results tables to show' in the OC Options? You can Redirect your results to your original image stack as well, just in case you are meauring intensities. As for the numbers, they are probably there...they're either too small or in a color that is similar to your object colors. Use a different LUT for your objects (e.g. glasbey on dark), The numbers are also in different slices of your z stack.

Hi @simoneceron1915. Perhaps the experts in the image.sc forum might be able to help with this: forum.image.sc/.

Hello @evanh1132. Thanks for watching. MorphoLibJ also has analysis tools. You can check out some of these tools here: ua-cam.com/video/YxCacx917tg/v-deo.htmlsi=Rh3OIxa8B4M-f8Pf

Hi @sanyamjain4210. Are your files large? If they are not really whole slide images, perhaps you can open your files in Fiji and use StarDist to segment the cells. Also, is the pixel size really that large? Perhaps there might be something wrong with the spatial calibration in your image.

Glad this helped, @SoleneVacher. Thanks for watching.

Hi @acooper6521. DiAna is more of a 3D image analysis tool (object-based co-localization and distance analysis). There might be a different plugin that caters better to what you want to measure in time lapse images.

Hello @spanishflea634. I have no experience with SFDI, so I can't give you a definite answer to your questions. Is it considered an optical image? If it is not, then I don't think deconvolution has any scientific validation for these types of images. Perhaps what you need is an enhancement of your image (although I can't be sure about this since I don't really know what your image looks like). If you could share your image, maybe I can suggest some methods to enhance its quality.

Hi @shruthib8176. I don't think TrackMate has analyzers that measure angular velocity. It does have features that track motility type and mean directional change.

@@johanna.m.dela-cruz Thank you very much for getting back to me. I used the xy coordinates from the TrackMate to calculate the angular velocity.

@@annisadwilestari4470 thanks for supporting my channel. I could try to do a tutorial like you suggested…I just need an image to work with.

Thank you so much for answering, it means a lot, because I'm currently struggle to learn how to do it. I'll be waiting for it.​@@johanna.m.dela-cruz

@annisadwilestari4470 do you perhaps have an image i can use? You can email it to me if you’re ok with me using your image.

@@johanna.m.dela-cruz hey I'm actually have not a picture from my self. Is it okay if I took it from Journal/someone research and send you?

@@annisadwilestari4470 I guess i just need to know what exactly I’m working with or how the image would look like (since this is not a field I’m used to).

Thank you Olivia (@oliviapietrobon5225) for your continued support. What type of measurements do you want to do on your IHC images?

@@johanna.m.dela-cruz Hello, I am analyzing images where I used immunohistochemistry to mark androgen receptors. Although it's not the same, when I saw your analysis in immunofluorescence, I noticed that the analysis is done for each ROI (Region of Interest) that you create. ¿Can I do the same in the case of immunohistochemistry? Another question, when you analyze each ROI, do you get a value for each one,To create the graph, ¿did you take an average of all those values? I hope my questions are clear. Thanks, Johanna!

@oliviapietrobon5225 As long as you arr able to segment the regions you want to measure, then, you can get measurements from each ROI. I am not sure what stain you used, but if you have DAB, quantifying intensity won’t be a sensible way to characterize antigen expression.

@@johanna.m.dela-cruz Hello, I use de tunel technique and caspasa 3 (for this i use DAB), now another question, I see that you're making a graph of the measurements; I'd like to know if you're calculating an average of the values from each ROI. My concern is that sometimes the values can be very different between each ROI, wouldn't that result in a high deviation?

@@oliviapietrobon5225 yes, it's an avaerage.

Hello @Lam-m3l. Thanks for watching my tutorial. Unfortunately, due to some permission issues, I can't share the images I used. However, there are several image datasets that you can use from this site: imagej.net/plugins/samples.

@@johanna.m.dela-cruz Thanks for the info. I was interested in the convoluted-looking fiber-like first dataset that you used. Not necessarily microglia though.

@Lam-m3l You’re right…that was a z-stack of microglia.

Hi @KritiAttri. On the right side of the Kappa console, you have to click on Curves and then select the curves you want to get their data exported.

@@peterlkk Split the channels and run deconvolution (and psf) on both channels individually.

Hi @jd9611. I don't believe Kappa has that feature.

Hi @tritoniadiomedea. The stacks used in this video were acquired as transmitted light images in a confocal microscope, so step size was constant. However, if images were acquired at various intervals (manual focus adjustment), EDF would still be useful.

and adding "wait" steps doesn't work

@kevinn1534 Perhaps there’s something missing in your code.

@@johanna.m.dela-cruz thanks for your reply... can I show it to you? I don't think is missing something, I literally use the macro recorder and I verify it.

@kevinn1534 you can email me. I’m not an expert in macros, but I can take a look at it.

@johanna.m.dela-cruz thanks a lot 🙏 Can you share me your email pls

@@sriti_hikari You can try Directionality or OrientationJ. My tutorial is here: ua-cam.com/video/tN0bEVsumJc/v-deo.htmlsi=vxeAMMU906Wsts0z

@@johanna.m.dela-cruz Hi thank you for your help- this will only give the orientation in the x-y-plane right? so the direction in z is not considered?

@@sriti_hikari Since each image in a z-stack is an optical section of your sample, fiber orientation shows the angle per slice. If you do a maximum projection of your z stack and then run the plugin, the orientation is based on the xy plane.

It’s heartwarming to receive your positive feedback. Thanks for your support.

@@usercvzu1256 You will first need to make sure you have a single file containing a sequence of your images. If you have several photos in a folder (File - Import - Image Sequence).

@@MDIqbalHossain-d8v hi! You can convert the label image stack into a binary stack by thresholding. Use a minimum threshold of 1.

Hi Zahra. If you go to my channel description, my email address should be there.

@@johanna.m.dela-cruz I can’t find there😅

@@zahraahmadi2234 If you are subscribed to my channel, under my name, you will find the statement "Listen, watch and learn..." click on more and you can view my email address. Sorry, I can't just advertise openly my email address. You would have to be on a laptop/computer (not phone) to see it.

@@johanna.m.dela-cruz I use phone and subscribe your awesome channel ( I don’t access a laptop or computer- would you please share your email address here)

@@johanna.m.dela-cruz I use phone and subscribe your awesome channel ( I don’t access a laptop or computer- would you please share your email address here)

@m35926 I think you can do this with a grid. It might be a tad more complicated, but check out this video about counting objects in a grid. ua-cam.com/video/fPRE8nePFgc/v-deo.htmlsi=4EET_DzS8yNJbDoM

Hi Octavia (@octaviasantisl). A labeled image/stack can be converted into a binary mask by using a low threshold of 1.

Hi Xander (@xandermolina5106). I don't really understand what you mean. If you have a circle, the curvature is calculated as the reciprocal of the radius.

@@TNTT1234 Hello! Thanks for watching. You can check out these videos for creating a stack from a sequence of images and doing a 3D reconstruction: ua-cam.com/video/_D1kbKjZtjk/v-deo.htmlsi=qNwIhuYjzaj0KEM- ua-cam.com/video/i-Ta4vCRW2s/v-deo.htmlsi=pHjihABbx4elO39E You can also try 3D Script: ua-cam.com/video/Zcb0hhZsYqU/v-deo.htmlsi=ZXyjb9FVpD2wLpR- For image registration, you can do this once you have a stack of images and before 3D rendering.

@@johanna.m.dela-cruz thank you so much. I am appreciated.

Hi Rupa (@rupachowdhury6164). If you use a larger gaussian blur filter, the resulting image will of course look more blurry. However, when you thrreshold the image from the subtraction and count masks (from Analyze Particles), the mask will just show the segmented objects. If you don't see the objects, change the LUT to glasbey-on-dark. You should be able to see the ROIs in the ROI Manager.

Hi @Luroog. Yes, this can be done. Are your objects moving over time or are the positions stable over time? If they are stable, you can check out this video: ua-cam.com/video/ENeCE_iur_o/v-deo.htmlsi=N-ppFZJCADbxSwj4

@@johanna.m.dela-cruz my objects are moving over time.

Hi Misha (@meesha5883). Thanks for your continued support. If you have your segmented cells added into the 3D ROI Manager, you can select these ROIs and apply them to your other channel, then do your measurements (just like you would on the original channel where the cells were segmented).

@@johanna.m.dela-cruz thank you for responding, I managed to apply the segmentation to another channel but how do I do thresholding for that channel (and add positive label for that marker to positive cells). Another problem is that the segmentation used nuclear stain and I want to dilate the ring of segment ring for each cell so that I can measure the marker that are expressed in the local cell area not just nucleus. I want to output a csv file with cells coordinates and info if they are positive for any of the markers.

@meesha5883 I believe you posted this question in the image.sc forum. For a label image stack, you can use MorphoLibJ. Go to Label > Label Edition. Dilate the objects (as many times as you like), then press done. You can then add the new label stack to the 3D Manager.

@@PatriciaNerydesiqueira Is your binary image with a white background? It might be that the objects are not really touching the edges of the image since there could be a row of white pixels between the particles and the edges. Try cropping the image and see if there is a difference.

@@johanna.m.dela-cruz Hiiii, Johanna, thanks for the quick response :D the image has a white background when I apply the tresholding, but I soon convert it to binary and the background turns black. but really, what you say about not touching the edges makes sense, but unfortunately there is no way for me to crop image by image, there are too many. and we are developing a macro to automate the process. do you have any other suggestions? *.*

@@PatriciaNerydesiqueira make sure the objects you are trying to measure have a value of 255. The background should be 0.

@@johanna.m.dela-cruz but if the image has already been binarized, is there any way to avoid having 0 and 255?

@PatriciaNerydesiqueira A binary image has 2 values - 0 and 255. You just need to make sure that the objects are 255. Just hover your mouse over an object and check the value displayed in the Fiji status bar.

Hola @jessyhell3541. Sorry I don't speak Spanish. You need to have an image open first before running the tool.

Hello Arpita. Are you doing analysis on a per-cell or per-mito basis? If you want results for each mitochondrion in each image, you will need per-mito calculation to identify mitochondrial sub-populations. Per-cell basis should give normalized values rather than raw totals.

@@johanna.m.dela-cruz I did both. In the control vs test I can see the changes, but when measuring the mean length or length/mito i didn't see any changes My another concern is also that if I change thresholding parameters in control or in test ami I not making the data biassed?

@@arpitadutta6737 The thresholding strategy of the plugin involves a local or adaptive thresholding algorithm that requires empirical determination of its settings. This means that finding optimal settings for thresholding values is largely based on your own observations. You will have to determine which thresholding algorithm to use based on your own judgement....Quantification will change based on the block size and C-values you use. S0, whatever method and parameters you choose will have to be applied to all your images (assuming they were acquired using the same settings).

@@johanna.m.dela-cruz heya. I tried to put adaptive threshold, changing the c value to 1 also didn't show any change, also I changed the block size into 0.25 from 1.125; that is the default. Still the thresholded image does not reciprocate to the original.

@@arpitadutta6737 Thresholding is always a challenge especially with complex images. I don’t know how your actual images look like, but you may want to try pre-processing them before using the plugin. Are your images 2D or are they z stacks? Some complex images might need deep learning algorithms for more accurate segmentation.

Hi Melania (@melaniabica-popi4279). Thank you for supporting my channel. With the right parameters and as long as the filaments are clear enough, I would think Ridge Detection will be able to measure these. You might have to do some pre-processing to get them to be more visible.

@@pacmorgado3443 it varies depending on the image. Your acquisition parameters may have been different from what the manual suggests. It usually is trial and error to get the “perceived” best block size. Thanks for supporting my channel.

@rikasilamdeen, On Trainable Weka Segmentation, the Plot result button generates a model performance chart (i.e. the ROC, precision/recall, etc. curves) or graphical plot based on the training data.These curves show the performance of the classifier based on the different thresholds that can be applied to the probability maps.

Hi @simonediana1958. Thanks for watching. I have not tried the Levan plugin, unfortuantely. I'm sure it serves its purpose in terms of image analysis. Good luck on your presentation.

@@johanna.m.dela-cruz thank you!

Thank you so much, Tay (@taystevens6009).

Hi @dr.sandeshkamdi4601. Yes, % overlap can be computed from the overlap area normalized to the area of either one of your objects (GAD or cFOS). Note, however, that the %overlap is an overall value. If you want to get individual overlap percentages, you can try DiAna (ua-cam.com/video/CBwWk7wP2kQ/v-deo.htmlsi=BOS8bozq90Gl4xAo).

Please check my reply in your other comment.