Three-dimensional intersection structure prediction for autonomous driving applications

A method comprising:

• applying, to a neural network, data representative of a two-dimensional (2D) space, the data generated using at least one sensor;
• computing, using the neural network and based at least in part on the data, a representation of three-dimensional (3D) world-space locations;
• converting the 3D world-space locations to the 2D space to generate 2D space locations;
• comparing the 2D space locations to 2D space ground truth locations associated with the sensor data;
• and
• updating one or more parameters of the neural network based at least in part on the comparing.

A processor comprising:

• one or more circuits to:
• apply, to a neural network, sensor data representative of an intersection in a field of view of at least one sensor;
• compute, using the neural network and based at least in part on the sensor data, data representative of three-dimensional (3D) world-space locations corresponding to line segments of the intersection;
• and
• based at least in part on the 3D world-space locations corresponding to the line segments, perform one or more operations associated with control of an autonomous machine.