Metrics
In this section, you can see statistics about your dataset. Currently, you can view the number of annotations and the number of objects for all datasets. You can also view the class distribution for datasets using an annotation taxonomy that contains objects and object instances.
The amount of Delivery Ready annotations within the selected dataset. An annotation is usually a set of shapes, drawn according to the defined taxonomy, on top of one unit of input data uploaded by you.
One unit of input data might result in multiple parallel quality-assured annotations, if this makes the combined annotation work more efficient and does not result in worse quality.
The amount of unique objects within the selected dataset. The uniqueness is valid per annotation, so inside an annotation of a video or across multiple data sources (cameras, lidars, ..).
An object in this context is mostly a 'thing' (car, person, traffic sign, ..) but can also be 'stuff' (road, debris, ..). That is, if a car is visible in two sequence annotations, it's one object in the first sequence and one object in the second sequence.
The number of instances across all objects and all sequences in the current dataset. One instance of an object is its appearance in one frame and one sensor.
Note! This information is only available for sequence datasets.
Number of frames across all sequences in the current dataset. Includes frames that don't have any annotations.
Note! This information is only available for sequence datasets.
The distribution of object classes within your selected dataset.
Note! This chart is only available for datasets using an annotation taxonomy that contains objects.
The distribution of object instance classes within your selected dataset.
Note! This chart is only available for datasets using an annotation taxonomy that contains objects and sequence inputs.
To see the exact amount of objects marked as a specific class you can hover over the specific class's bar. A tooltip will then appear which shows the amount of objects of the specific class.