Automated labelling
Automated labelling is a way for you to batch-process images in your project. Essentially, you take a model used in your project (Object Detection Assistant Instance Segmentation Assistant or SES) and apply it to either all images that haven't been annotated in a project or a specific part of it.
The feature costs credits to use. The cost is estimated at the bottom of the screen and will be finalized at the end when all the labels are created. You will be able to work on your project while Hasty annotates the images for you, but there might be some strangeness going on with images suddenly getting annotations while you are doing your work. You will be notified once the labeling is completed.
Important: This feature is only enabled when we are sure that the accuracy of annotations generated by the feature will be acceptable. We enable auto-labelling when both user behavior and machine learning metrics indicate that a high level of automated accuracy has been reached. To get to this point, you will need to annotate anywhere from 500 to 5000 images. For some very large projects with very diverse data, that number might be even larger.
If you want to try the feature but it hasn't been made available yet, contact Herbert ([email protected])
Copy link