In 3D Cuboid Annotation, 2D images are taken and annotated with 3D models. Consequently, various elements like height, width, length, depth of the image are added. This enables the computer to capture the precise dimensions of an object.
Capabilities
3D Bounding Box Use Cases:
Camera based perception in autonomous vehicles
Used to train computer vision models for spatial cognition from 2D images or videos. Relative distance of each mobile object from the ego car and vanishing point can be measured.