Overview

Each time you change something that impacts your raw data images, or how MyoVision will analyze the cell boundaries, you need to create a new classifier. Examples of reasons to do this include:

• changing a setting on the microscope
• using a different stain
• working with a different type of muscle (for example, plantaris verus soleus, or human versus mouse)
• changing percent_saturation in <image_to_label_parameters> in a configuration file

Creating the classifier, involves three steps:

• Segment a test image into test blobs
• Manually classify each blob as one of:
  • a plausible fiber (code 1)
  • potentially connected fibers (code 2)
  • something else, for example, interstitial space (code 3)
• Training the classifier

This page leads you through an example workflow.

Workflow

• Segment a test image into test blobs
• Manually classify blobs
• Train the classifier

Once you have created a new classifier, it’s a good idea to test it.

• Testing your new classifier