The Anomaly Detection tool in Aurora Focus uses the Aurora Deep Learning Editor.
Before running the editor, provide the folder where the created models are stored. The folder you create is the model's name in Aurora Focus.
In Aurora Focus, click
Create Model
to launch the editor. After running the editor, you can create a new model or edit an existing one.
If an existing model is not shown in Aurora Focus, select it from the drop-down and click
Edit
.
In Deep Learning Editor, you can use an existing model or create a new one by selecting a model type from the drop-down menu.
The upper left corner of the home screen provides tools for training images. Click
Add
to add new images to the Editor.
Images added using
Toggle Class
and
Select type
can be marked as ready to be trained on.
Images with a model are marked as training with a green checkmark.
In some cases, it may be helpful to include irregular images marked as test images to train the model.
The
Training Parameters
to decrease or increase the complexity and downscaling of the model. Both contribute to the training time and the working time of the model. The effect of enabling these settings is observable by clicking
Pre-processing
above the previewed image.
After the settings are configured, click
Run
to start generating the model.
If the Deep Learning Editor is being run for the first time, the first click launches the Deep Learning Service. The second starts generating the model.
