Cell classification

Описание: Sara McArdle presents how to train an object classifier.
Files https://drive.google.com/drive/folder... and a related presentation https://docs.google.com/presentation/... are available for download.

00:00 Cell classification workflow
00:06 Introduction to Lung1 HE image by Ken Kim
04:16 Doing stain deconvolution
06:14 Inspecting residual channel
06:59 Starting cell detection
07:27 Adjusting cell visualization
07:54 Fine-tuning cell detection
08:40 Need help? Ask a pathologist!
09:25 Using context help
09:58 Doing cell detection on the whole section
11:02 Adding measurements
11:37 Measuring textures in cells
12:13 Adding smoothed features
12:58 The order of adding features is important
13:59 Adding annotations for the classifier
16:09 Working with the Object classifier
18:24 Improving the classifier
20:17 Balancing the training set
21:49 Saving your training data for future
23:05 Starting with a clean slate
24:10 Training an eosinophil classifier
24:35 Using the points tool for annotating rare training objects
28:38 Examples of eosinophils and false-positives
32:01 The importance of not confusing the classifier
33:10 Hiding unclassified objects
34:24 Different classifier algorithms
35:20 Calculate variable importance
37:02 Changing the features used for training
40:42 Training the classifier across different images
43:29 How to deal with classifying different objects
43:58 How to classify non-cellular objects?

If you have any questions about the videos or the training course, please ask on this image.sc forum post. https://forum.image.sc/t/qupath-train...
The more information you can include about your question, the more helpful our answers can be.

Please follow this document to get everything you need (software, files, cheat sheets, and additional resources):
https://docs.google.com/document/d/1V...