Enhanced 3D Bounding Boxes Object Annotations for AI Training
January 2021
Combining the realism of our synthetic data with the flexibility and power
of our cloud platform, we’ve taken our annotation capabilities to the next
level by separating the elements within a specific asset. As in the image, it is
important to distinguish between the motorcycle and its rider as two
independent objects. By doing so, we are enabling the creation of separate
bounding boxes around those sub-elements.
With this new Cognata feature, each sub-assembly object is now considered
independently in the ROI and Object Annotation files for improved ground
truth. Experience training data comprising some of the world’s most
extensive asset catalogs annotated with 3D bounding boxes for ground
truth with accuracy and detail like none other.