Process flow for image analysis using AI

Cell Pocket

Add Bookmark

STEP 1. Set data analysis goals

Clearly identify.
Example: Evaluating the variability of the spheroid area and circularity

Set data analysis goals

 

 

STEP 2. Acquire AI training data

Acquire images necessary for obtaining desired results.
Verification is possible from a minimum of about 10 image data sets.

Acquire AI training data

 

 

STEP 3. Train AI model and assess performance


Train the AI model and assess its performance with test images. Cell Pocket™ automatically assigns a test image, numerically evaluates the entire test image, and indicates any regions of the test image where errors were predicted.
The figures below are used to assess whether spheroid regions were correctly predicted.

Train AI model and assess performance

 

 

Note:  Processes for calculating area and circularity values from predicted spheroid regions are specified in data analysis recipes.
            In this example, the results confirm whether spheroid regions can be correctly identified.