Description

You can use Nucleus to find missing annotations in your dataset using high confidence predictions from your models to resend them for annotation.

Pre-reqs: Dataset, Annotations, Predictions

Steps

1. Open a dataset with annotations & predictions uploaded
2. Go to the objects tab and open the search options sidebar
3. Under the model filter select the model you want to use
4. Under the object filter select false positives i.e. (prediction & annotation mismatch)
5. Set the IOU slider to 0 to get objects with no annotations
6. Sort the results by confidence in descending order to get high confidence predictions

Among the final results you can see that many of the predictions are clearly right. However, these are classified as false positives because the annotator missed the label. You can compile these high confidence predictions in a slice to resend them to annotation.