How the behaviors are classified

Behavior annotation in SmartAnnotator consists of the following steps:

1. The top view videos are preliminary annotated with the Rat Behavior Recognition Module or Mouse Behavior Recognition Module of Ethovision XT version ***
   See Behavior Recognition Requirements for the best practices for camera setup, arena setup and detection settings.
2. Upon selection of the behaviors to annotate, SmartAnnotator divides the videos into fragments, each with a separate event, based on the automatic behavior classification in EthoVision XT
3. The behaviors for which the model expects to get the most valuable information are shown to the user, to manually annotate. 
4. The AI model is updated with these manual annotations. 
5. Steps 3 and 4 are repeated until the classifier reaches an accuracy threshold.
6.  The resulting classifier is applied to all unlabeled events, giving these an automatic label if the model is certain enough.
7. The AI model picks a number of unlabeled events that are most informative for the classifier, and gives them to the user for manual labeling