Can anyone share a raw video data for creating whole-slide image?

Here you can suggest and discuss Websites, books, and other resource materials.
Post Reply
Message
Author
Spiso
Posts: 3
Joined: Fri May 29, 2020 7:51 am

Can anyone share a raw video data for creating whole-slide image?

#1 Post by Spiso » Fri May 29, 2020 8:24 am

Hello everyone,

I am looking for video data to download. My task is to create a software to stitch the video into a whole-slide image. So the data I am looking for has to be static, with only the camera moving horizontally and vertically to slowly cover the whole object. Software that does this is for instance MS ICE or microvisioneer, but I have a task to create my own.

Can anyone upload their data for me? Or do you know of any sources? I could easily find stitched whole-slide images in .svs or other formats, but nowhere on the internet I could find the videos used to create these images, which is what I need.

Thank you very much for sharing any information with me.
Spiso

microb
Posts: 729
Joined: Wed Mar 27, 2019 6:39 am

Re: Can anyone share a raw video data for creating whole-slide image?

#2 Post by microb » Fri May 29, 2020 3:52 pm

Any such video would probably involve some focusing, so feature points will be at different layers of a specimen slide. Maybe as the video runs, one section is filmed, that previous data can match up feature points to the next exposed section of the slide as it racks focus again. But you'll have feature points blur in and out, you'll need to find that mapping of feature points at one z to its neighbors so you can focus stack the slide -- or at least continue the one layer you want. Ideally, you would also have micrometer data on the Z, and optionally XY encoders too. But you could visually track the XY and optionally scale it accurately if a calibration slide image was given. The Z distance changes would allow you to create volume data stitched together. You could create something like a VDB file for example, source is on github.

Do note, the microscope is not like CNC material or a CNC table. The sample in the slide will not be perfectly level when it comes to focus, the slide itself may not be perfectly level, and the table as it moves has probably never been calibrated since the factory so it will not be level. So some focusing during a video will be require going across a slide.

Spiso
Posts: 3
Joined: Fri May 29, 2020 7:51 am

Re: Can anyone share a raw video data for creating whole-slide image?

#3 Post by Spiso » Mon Jun 01, 2020 8:47 am

Hi,

thank you for the explanation. So I have to count with the fact that at some point it will be refocused. I believe I can deal with that, but for now I really need a video data file.

I can see in this instructions video https://youtu.be/XipXRo7iftg?t=393 that he has uploaded a video data to MS-ICE. I am looking for any such data to play with. Do you have any that you could upload for me?

Thank you.
Spiso

User avatar
mrsonchus
Posts: 4175
Joined: Tue Feb 03, 2015 9:42 pm
Location: Cumbria, UK

Re: Can anyone share a raw video data for creating whole-slide image?

#4 Post by mrsonchus » Mon Jun 01, 2020 10:56 pm

Hi, here's an example I made back in 2016 of a video stitched with MS-ICE.
The final stitched image is at the end of this edited/annotated video...


Here's a video-scan I took this evening of part of a microscope-slide (sectioned at 5µ and stained with safranin + fast-green stains) Cat's Ear 'Hypochaeris radicata' weed from my garden, made a few days ago.
This is the video - unedited -


This is the result of this video being used with MS-ICE to create a stitch,
Result of video-stitch with MS-ICE
Result of video-stitch with MS-ICE
video 01_06_20_0010_stitch.jpg (109.9 KiB) Viewed 4022 times
John B

Spiso
Posts: 3
Joined: Fri May 29, 2020 7:51 am

Re: Can anyone share a raw video data for creating whole-slide image?

#5 Post by Spiso » Wed Jun 03, 2020 7:32 am

Hi,

yes, this is what I was looking for. Thank you very much!

Spiso

Post Reply