Path 1: AI-assisted analysis
Overview
AI-assisted analysis is the fastest way to get to quantitative results for small image sets. However, as there is still a manual-edited portion, it may not be suitable for analyses that have too many objects, such as:
Large batches of images with many objects per image
Segmenting out small objects like cells on whole-slide images
Tutorial
1. Upload images
If you already have uploaded files, you can skip this step.
Navigate to Files from the left sidebar, and press Upload. You'll be able to upload from your computer, or import from a third party (even S3). This will take a second to upload and process into files.
2. Start the analysis
Once your files are uploaded, go to AI Analysis on the left sidebar, and click Start a new analysis run. Then choose AI-assisted, and choose the files that you uploaded. Then press Start. This should open you into an AI-assisted analysis window.
3. Label
Your labeling tools and how to use them
Select your first class on the left toolbar. Once you click, you will have three different tools you can toggle between by clicking them: AI Detect, AI Select, and Pen. You'll also now be in drawing mode, so you will now need to hold Space to be able to pan across your image.
AI Detect is the fastest, and you should use it unless it isn't working well for you, filling in the gaps with AI Select. The Pen is the slowest, and should be used if AI Select is not working for you. AI Detect is only applicable to object classes.
To use AI Detect on a tile, drag a tight box around one example of your object, using the crosshairs to help you. AI will locate the position of similar-looking objects, and also segment them.
Using the filters
After you use AI detect, you'll need to filter the results, using the sliders. The confidence score slider controls the model score threshold at which objects are filtered out. Drag higher to get fewer labels, and lower to get more labels.
The size range slider has a minimum and maximum upper bound. You can filter out or include larger or smaller objects using these ranges.
To use AI select, "prep" a tile by clicking on it, which should be ready in around one second. You will start seeing predictions in real time when hovering.
Labeling objects with AI
Label your object by clicking, or drag a box around it for a more precise label on difficult objects.
Refine by clicking. Click outside the object to grow your object, and inside to shrink your object.
Press A on the keyboard to start a new object.
Drag and click around your object to get your label.
Select
To select one object, click on it on the image, or select it from Labeled objects on the left. To select many, hold Ctrl or Cmd and drag on the image.
Delete
To delete selected objects, you can press Backspace. Press Enter to confirm.
Edit
Make sure that you are out of drawing mode (press ESC). Select an object by clicking it in the viewer or from the Labeled objects section. Then, press the key e or click on the pencil in the Labeled objects section to enter editing mode. From here, you can choose to Add or Subtract from the object. Draw an object that will be added or subtracted from that object. Once your object has been completed, your changes will be automatically saved.
Other settings you might use for labeling
You might need to switch images, add images, configure classes, or adjust your brightness/contrast settings. Find the instructions for those below.
Changing class color
You can also choose from different colors to represent the class. Simply click the colored circle to open up a color picker.
Changing the name or deleting a class
Hover over the class to reveal the 3 dots button, which you can click. From there, use the dropdown to rename or delete a class. If you delete a class, every object created for that class will be deleted as well.
Add a new class
Click the + button next to your size group on the left sidebar to add a new class. You might need to add a new size group if your new class represents an object that is of significantly different size or a different segmentation type.
Adding a new size group
Click Add size group from the left bar, and follow the setup flow to select your segmentation type and tile size.
What is a size group and why should I make one?
A size group is a group of classes that use the same tile size and segmentation type. They allow your AI model to see your objects at different scales in the same image, to optimize your accuracy. This also allows you to infer interesting parent-child hierarchy information from very different sized objects.
Create a new size group if you want to:
Blending region segmentations (confluence, inflammation, tissue regions) with object (cell, cyst, organoid) in one model.
Blend two very different sized objects in one model (organoid and cells, tumor and immune cell, crypts and goblet cells)
Separate labeling for sparse vs. dense classes (rare cell type vs. very common cell types)
Keep in mind that adding a new size group increases training and analysis credit usage.
Switching between images
Click on the dropdown under Choose image on the left sidebar, and select a new image to change to it. You can scroll, filter, and search to find your image. Alternatively, press Next or Prev to move forward or backwards.
Adding images
If you need to upload more images, you'll need to exit the dashboard entirely and add more images. Click the X in the top left to exit, and look for the Add or remove data tab.
4. Quantify your labeled objects
Submit for analysis
Once you've labeled all of your objects of interest, you can quantify them. Don't worry, you'll be able to add more objects later and re-quantify.
Click the Quantify objects button, to open up your configuration window.
<Image of choose metrics + hierarchies>
On this window, you can choose the metrics you'd like to quantify, including channel intensities and morphological characteristics. Choosing more metrics can lead to slower processing. Then, press submit.
Interpreting your results
You're done and ready to look at your results! View the Results dashboard overview to understand this page.
Last updated