View source for Nvidia corporation (20240273926). USING NEURAL NETWORKS FOR 3D SURFACE STRUCTURE ESTIMATION BASED ON REAL-WORLD DATA FOR AUTONOMOUS SYSTEMS AND APPLICATIONS simplified abstract

=USING NEURAL NETWORKS FOR 3D SURFACE STRUCTURE ESTIMATION BASED ON REAL-WORLD DATA FOR AUTONOMOUS SYSTEMS AND APPLICATIONS=


==Organization Name==

[[:Category:nvidia corporation|nvidia corporation]]


[[Category:nvidia corporation]]

==Inventor(s)==

[[:Category:Kang Wang of Bellevue WA (US)|Kang Wang of Bellevue WA (US)]][[Category:Kang Wang of Bellevue WA (US)]]

[[:Category:Yue Wu of Mountain View CA (US)|Yue Wu of Mountain View CA (US)]][[Category:Yue Wu of Mountain View CA (US)]]

[[:Category:Minwoo Park of Saratoga CA (US)|Minwoo Park of Saratoga CA (US)]][[Category:Minwoo Park of Saratoga CA (US)]]

[[:Category:Gang Pan of Fremont CA (US)|Gang Pan of Fremont CA (US)]][[Category:Gang Pan of Fremont CA (US)]]

==USING NEURAL NETWORKS FOR 3D SURFACE STRUCTURE ESTIMATION BASED ON REAL-WORLD DATA FOR AUTONOMOUS SYSTEMS AND APPLICATIONS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 20240273926 titled 'USING NEURAL NETWORKS FOR 3D SURFACE STRUCTURE ESTIMATION BASED ON REAL-WORLD DATA FOR AUTONOMOUS SYSTEMS AND APPLICATIONS

'''Simplified Explanation''': 

The patent application discusses the training of a deep neural network to predict a dense representation of a 3D surface structure using real-world data collected by vehicles equipped with image and LiDAR sensors.

* '''Key Features and Innovation''':
    * Generation of training dataset from real-world data collected by vehicles.
    * Estimation of a sparse representation of a 3D surface structure from image data.
    * Smoothing and alignment of LiDAR data to generate ground truth training data.
    * Projection of identified 3D points to create a dense representation of the 3D surface structure.

'''Potential Applications''':
This technology can be applied in autonomous driving systems, robotics, augmented reality, and urban planning for accurate 3D surface structure prediction.

'''Problems Solved''':
The technology addresses the challenge of accurately predicting dense representations of 3D surface structures from real-world data, which is crucial for various applications such as autonomous navigation.

'''Benefits''':
The benefits of this technology include improved accuracy in 3D surface structure prediction, enhanced performance of autonomous systems, and better decision-making capabilities in various industries.

'''Commercial Applications''':
Potential commercial applications include autonomous vehicles, construction and infrastructure development, urban mapping, and virtual reality systems.

'''Prior Art''':
Researchers can explore prior art related to LiDAR data processing, image data analysis, and deep neural network training for 3D surface structure prediction.

'''Frequently Updated Research''':
Stay updated on advancements in LiDAR technology, image processing algorithms, and deep learning techniques for 3D surface reconstruction.

'''Questions about 3D Surface Structure Prediction''':
1. How does the technology of predicting dense representations of 3D surface structures benefit autonomous driving systems?
2. What are the key challenges in accurately estimating 3D surface structures from real-world data?


==Original Abstract Submitted==

in various examples, to support training a deep neural network (dnn) to predict a dense representation of a 3d surface structure of interest, a training dataset is generated from real-world data. for example, one or more vehicles may collect image data and lidar data while navigating through a real-world environment. to generate input training data, 3d surface structure estimation may be performed on captured image data to generate a sparse representation of a 3d surface structure of interest (e.g., a 3d road surface). to generate corresponding ground truth training data, captured lidar data may be smoothed, subject to outlier removal, subject to triangulation to filling missing values, accumulated from multiple lidar sensors, aligned with corresponding frames of image data, and/or annotated to identify 3d points on the 3d surface of interest, and the identified 3d points may be projected to generate a dense representation of the 3d surface structure.



[[Category:G06V20/64]]
[[Category:B60G17/0165]]
[[Category:B60K31/00]]
[[Category:B60W60/00]]
[[Category:G01S17/89]]
[[Category:G01S17/931]]
[[Category:G06F18/214]]
[[Category:G06V20/58]]
[[Category:CPC_G06V20/64]]