Synthetic 3D Object Generation from Incomplete Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in generating synthetic three-dimensional objects for simulated environments due to incomplete sensor data, which limits the creation of digital twins for novel objects encountered during operations.
Innovation Solution
The use of feature vectors from sensor data, processed by machine-learning networks, to identify and modify existing 3D object references in a database, generating complete synthetic objects by adjusting geometry and color information, and storing these new objects for future use.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If sensor data is used to generate synthetic 3D objects, then the library of objects in the database is expanded, but the sensor data is incomplete which limits the creation of complete digital twins
Solution Approach 1:
The patent creates complete synthetic 3D objects by copying and modifying existing complete 3D objects from the database. When novel objects are detected with incomplete sensor data, the system finds similar existing objects in the database, copies their complete geometry and color information, and modifies them to match the detected object's characteristics, thereby creating complete digital twins from incomplete data.
Solution Approach 2:
The system changes parameters of existing 3D objects to generate new synthetic objects. By extracting features from incomplete sensor data and comparing them with database objects, the system modifies geometric parameters, color parameters, and other attributes of reference objects to create accurate representations of novel objects while maintaining complete 3D structure.
2Manufacturing precision
If feature vectors are used to query the database for similar objects, then complete synthetic objects can be generated, but the processing complexity increases
Solution Approach 1:
The patent replaces complex manual or traditional 3D modeling processes with automated machine learning networks. These networks automatically extract features from sensor data, query the database using feature vectors, retrieve similar objects, and generate synthetic 3D objects through automated parameter modification, eliminating the need for manual modeling while maintaining high accuracy.
Solution Approach 2:
The system performs self-service by automatically completing the 3D object generation process. When incomplete sensor data is received, the system autonomously queries the database, identifies suitable reference objects, extracts their complete geometry and color information, and generates the final synthetic object without human intervention, thereby managing complexity through automation.
Data Source
AI summary
The disclosed technology provides solutions for generating synthetic objects and in particular, for generating three-dimensional objects using known geometries that can be queried from a database using feature vector information. A process of the disclosed technology can include steps for receiving sensor data including a detected object, wherein the sensor data represents one or more three-dimensional (3D) attributes of the detected object, processing the sensor data to identify at least one intrinsic attribute associated with the detected object, and querying an object database, based on the at least one intrinsic attribute, to identify a stored object that is similar to the detected object. In some aspects, the process can further include steps for modifying one or more attributes of the stored object based on the one or more 3D attributes of the detected object to generate a synthetic object. Systems and machine-readable media are also provided.