Nvidia corporation (20240320923). VIEWPOINT-ADAPTIVE PERCEPTION FOR AUTONOMOUS MACHINES AND APPLICATIONS USING SIMULATED SENSOR DATA simplified abstract
Contents
VIEWPOINT-ADAPTIVE PERCEPTION FOR AUTONOMOUS MACHINES AND APPLICATIONS USING SIMULATED SENSOR DATA
Organization Name
Inventor(s)
Tae Eun Choe of Belmont CA (US)
Minwoo Park of Saratoga CA (US)
Jung Seock Joo of Los Altos CA (US)
VIEWPOINT-ADAPTIVE PERCEPTION FOR AUTONOMOUS MACHINES AND APPLICATIONS USING SIMULATED SENSOR DATA - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240320923 titled 'VIEWPOINT-ADAPTIVE PERCEPTION FOR AUTONOMOUS MACHINES AND APPLICATIONS USING SIMULATED SENSOR DATA
Simplified Explanation: The patent application discusses systems and methods for adapting perception in autonomous machines using a 3D perception network trained with simulated data to handle unavailable target rig data.
- The 3D perception network is trained as part of a training network using simulated source and target rig data.
- A consistency loss is used to minimize differences in transformed feature maps extracted from simulated data.
- The paths through the training network are designated as the 3D perception network, which can then perform perception tasks using target rig data.
Key Features and Innovation:
- Adapted perception for autonomous machines
- Training a 3D perception network with simulated data
- Consistency loss to minimize differences in feature maps
- Designating paths in the training network as the 3D perception network
Potential Applications:
- Autonomous vehicles
- Robotics
- Surveillance systems
Problems Solved:
- Handling unavailable target rig data
- Improving perception accuracy in autonomous machines
Benefits:
- Enhanced perception capabilities
- Improved performance in autonomous systems
- Adaptability to varying data availability
Commercial Applications: Adapted Perception Technology for Autonomous Machines: Enhancing Perception Accuracy and Adaptability
Questions about Adapted Perception for Autonomous Machines: 1. How does training the 3D perception network with simulated data improve its performance? 2. What are the potential challenges in implementing this technology in real-world autonomous systems?
Original Abstract Submitted
systems and methods are disclosed relating to viewpoint adapted perception for autonomous machines and applications. a 3d perception network may be adapted to handle unavailable target rig data by training the one or more layers of the 3d perception network as part of a training network using simulated source and target rig data. a consistency loss that compares (e.g., top-down) transformed feature maps extracted from simulated source and target rig data may be used to minimize differences across training channels. as such, one or more of the paths through the training network(s) may be designated as the 3d perception network, and target rig data may be applied to the 3d perception network to perform one or more perception tasks.
