Dataset Insights
Last updated
Was this helpful?
Last updated
Was this helpful?
In "Preview", click "Insights"
You will find statistics, metrics and insights which help you to manage and have an overview look of your dataset, scroll down to see more.
Overview will give you a look of total annotations and issues created.
You can also check daily change and average annotations/issues per day. This insight will help you to check the productivity of your team.
In overall progress, you will see the number of data in each status: "Not Started", "In Progress" and "Completed". This will be helpful in tracking the progress.
With this part, you can check quality insights, which includes:
Precision score: how accurate the annotations (by the annotators) are, as compared to the reviewer(s). High precision means most annotations by the annotators are correct.
Recall score: how well the annotators are at identifying the objects of interest. High recall means most of the objects are correctly identified, whereas low recall means most of the objects are missing.
Issues/Annotations: the smaller of this metrics is, the better the annotation be.
These two metrics can give you a more detailed look at our member's performance.
You can check through time by changing time scale, the number of annotation/issue made will be shown here too.
This one will count the number of each annotation appears in your dataset.
As its name, this one will show you the average annotation per image of the whole dataset.
Correlation shows how frequently a pair of labels occurs together in the same image.
