AI Pixel Classifier: How to extract quantitative information from an image
Learn how to extract quantitative data using AI pixel classification.
More about Mica: https://fcld.ly/mica-yt-tut.
# Transcript ############
Mica simplifies your workflows radically, but you are not finished until you extract quantitative information from the image.
Here’s a quick tutorial on how to easily extract quantitative data with Mica.
First record your multicolor image.
In this example, we will count the nuclei we can see on that image.
Click on Learn and then load the image you are interested in.
You have two classes: the nuclei and the background.
Draw a background area first.
Second, draw the subject of interest.
Mica will then create a preview for the annotation you created.
You can do the whole training if you’re happy.
You have now trained an AI model using pixel classification to segment your nuclei.
You can also use this model for future experiments.
You can access all of the models you have trained by going to Results and selecting the image you want to quantify. Then, click on Analysis.
Choose the one you’re interested in, and click Start.
The data can be displayed as scatterplots, histograms or boxplots.
# Mica ############
Mica is the first fully integrated imaging Microhub in the world. It seamlessly integrates widefield imaging, AI-supported analysis, and confocal image. All in one incubator that protects samples.
Researchers of all levels can work on a single, easy-to-use digital imaging platform. They can move confidently from the setup phase to the beautifully visualized result, providing true access for everyone.
Users can now capture four times more data in a single exposure using either widefield or concentric imaging. With a single click, they can switch between widefield and confocal modes without moving the sample.
Source: