Pete Bankhead Profile picture
Aug 19, 2020 21 tweets 9 min read Read on X
I say 'almost all', because I need to mention the command list early: with 'Ctrl + L' you get a searchable list of everything in the menus.

Having told you that, I can now ignore the menus & focus on shortcuts - safe in the knowledge you can find things if you need them. 2/20
(If you *really* like the command list, you can turn it into a command bar and give it a special place at the top of the viewer... but I probably wouldn't unless I broke my 'L' key.) 3/20
Anyhow, annotations.

QuPath's annotation tools hang out on the top left of the toolbar & all have quick shortcuts to activate them. 4/20 Image
*Most* of the time, you should have the 'Move' tool selected (to avoid annotating by accident).

Beside it live the rectangle, ellipse & line tools to draw... well, rectangles, ellipses & lines.

The 'Shift' key constrains their shapes if needed. 5/20
But there are more tools, with lots of tricks & shortcuts to use them efficiently.

For example, there are polygon & polyline tools - which switch to become freehand tools if you feel like dragging the mouse for a bit... 6/20
...a zoom-dependent brush (which quickly toggles to an eraser with the 'Alt' key)... 7/20
...a color-sensitive wand (also zoom-dependent, can become an eraser & incorporates any color transforms)... 8/20
...and optional snapping to avoid overlaps & help create dense annotations (press Ctrl + Shift when using the brush or wand).

These tricks are part of the reason QuPath is often used for #deeplearning annotation. 9/20
Toggle the display of annotations by pressing 'A' (or the toolbar button), or whether they are filled with 'Shift + F'.

If an annotation is selected, it remains displayed even when the others are hidden. 10/20
If you're not sure what other options you have, bring up the 'Command list' & start typing.

This helps find commands to duplicate, merge, split, fill or expand annotations.

Note: You can contract too - just use a negative expansion. 11/20
Once you know what's there, you can get creative.

For example, expand a line with 'Remove interior' selected & then split the result to get bands of a fixed width inside & out (perhaps after cleaning up the ends a little bit).

Handy for tumor margins, for example. 12/20
Under the 'Annotations' tab, there's a list of all the annotations for the current image.

Measurements for the current selected annotation are shown at the bottom. 13/20 Image
Tip: press Enter with an annotation selected in the viewer to set its name & color*.

You can also 'lock' annotations to avoid accidentally editing them. Locked annotations can still be deleted (press 'Backspace').

*Or right-click it in the list & choose 'Set properties' 14/20
Names are displayed on the image (hide them by pressing 'N'), and descriptions appear as tooltips. 15/20
By default, selected annotations are in yellow.

If you want to select more than one, you can:
* choose them in the list
* click with the 'Move' tool selected & the 'Alt' key pressed
* turn on 'Selection mode' with the big 'S' button and draw around the ones you want 16/20
I say 'by default', because lots of QuPath is customizable. Click the cog wheel to change the preferences... using the search bar to find what you need. 17/20
When you're done, choose 'File -> Save' (or Ctrl + S, or agree when QuPath asks) & your annotations are saved.

No need to specify where: because you definitely followed the tip in the last tweetorial (right?) QuPath will store them in your project. 18/20

That should be all you need to keep your annotations inside QuPath.

If you need to get them *out*, there's a whole section in the docs describing ways to export them... as GeoJSON, WKB, WKT, binary images or labelled images. 19/20

qupath.readthedocs.io/en/latest/docs… Image
That's all for now. Next time we'll get into detecting & classifying things. 20/20

qupath.github.io

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Pete Bankhead

Pete Bankhead Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @petebankhead

Sep 11
Let me introduce: InstanSeg 🦠🔬💻👩‍🔬

This *would* have been a short thread about Thibaut Goldsborough’s PhD work… but he solved too many problems.

Now it's a long thread about 2 preprints, a whole new approach to cell segmentation & #opensource software to make it easy to use
Preprint #1: segmenting nuclei in pathology images.

Some methods exist - but it remains a hard problem.

Partly because the images are complex & variable, partly because they are *huge*.

Here’s a 38 GB OpenSlide image. Nuclei are small & there are lots.
doi.org/10.48550/arXiv…



Image
Image
Image
Image
Ideally, you’d have a method that’s fast, accurate, open, user-friendly & runs in the software you want to use - on the computer you’ve got.

Achieving all of those is hard. All methods make tradeoffs. Image
Read 20 tweets
Dec 3, 2022
QuPath v0.4.0 is now available!

Download it at qupath.github.io

#opensource #bioimageanalysis #digitalpathology #java #javafx QuPath homepage
There are so many new things that it'll take some days to describe even just the main ones.

For now, I'll start by mentioning a few of the user interface improvements.

QuPath v0.4.0 is more welcoming than previous versions... and a bit more stylish.
In fact, you can even style it in your own unique way if you really want to, thanks to #javafx & css (ideally in a nicer way than in my screenshots).

qupath.readthedocs.io/en/0.4/docs/re… ImageImage
Read 31 tweets
Apr 20, 2022
If you want to learn about #bioimageanalysis I've written a free & open textbook that tries to help:
bioimagebook.github.io

Thanks to the wonder of @ExecutableBooks & other modern magic it's not quite like a normal book... (thread)

@OpenEdEdinburgh @NEUBIAS_COST @BioimagingNA
First, the book tries to cover the main concepts, independently of any software, in a practical way.

This includes common pitfalls & problems, like data clipping, that can doom analysis from the start (2/n)
It also includes tricky stuff important for a lot of microscopy image analysis, like noise distributions & the signal-to-noise ratio... (3/n)
Read 15 tweets
Feb 5, 2021
A small number of people know the real background story to @QuPath, but most don't.

I didn't plan to ever tell it publicly, until a Google Alert today caught my eye.

A thread about open science & academia 👇 (1/n)
The short version is that I single-handedly wrote the software as a postdoc but was blocked from releasing it open-source for years, while the environment in which I was working became increasingly toxic.

I handed in my notice as a last-ditch attempt to see it released. (2/n)
This worked - but meant I was out of academia, and my old group were free to take the credit.

Which they did.

It was strange to see people suddenly become huge fans of open science, speaking like they were my biggest supporters rather than the reason I left. (3/n)
Read 10 tweets
Aug 20, 2020
@QuPath #tweetorial 3!

This will build on the last one & take us into cell detection, measurement & classification... with some extra visualization tricks along the way.

There will be some new things from v0.2 too. 1/26

Let's jump in with a familiar example: calculating the % of cells positive for Ki67 (i.e. brown).

To do this:
* Annotate an area of interest
* Run 'Positive cell detection'

QuPath detects the cells & provides the result, often in a matter of seconds. 2/26
In practice, you might want/need to tweak the settings to get better results.

There's a whole YouTube tutorial about that youtube.com/playlist?list=… )

There's also a publication independently comparing QuPath & other software for Ki67 doi.org/10.1038/s41374… 3/26
Read 26 tweets
Aug 18, 2020
Time for a quick introduction to @QuPath in tweetorial form.

First up: a little background on how QuPath differs from other great #opensource tools for #bioimageanalysis like @FijiSc & @CellProfiler - and how to get started viewing images. 1/12

qupath.github.io
QuPath's most obvious distinguishing feature is that it handles whole slide images. These are ultra-large 2D images, often up to 50 GB in size.

Whole slide images are everywhere in #digitalpathology & increasingly common in research. 2/12
A single whole slide image can be more than 200k x 100k pixels in size & contain a huge amount of information that matters to researchers & clinicians.

The trouble is trying to wring that information out of billions of pixels. 3/12
Read 12 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(