Wheel Performance Prediction From 2D Images Without 3D Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for predicting wheel performance in vehicles require converting 2D wheel images to 3D images, which degrades their usefulness and efficiency.
Innovation Solution
An apparatus and method using a convolutional autoencoder (CAE) to generate a latent space for 2D wheel images, extract and learn performance values, and predict performance without converting images to 3D, employing transfer learning and generative design to enhance dataset generation and analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D wheel images are converted to 3D images to predict wheel performance, then prediction accuracy is improved, but processing complexity and time consumption increase
Solution Approach 1:
The patent extracts and utilizes only the essential 2D image data directly, eliminating the unnecessary 3D conversion process. By focusing on extracting performance-relevant features from 2D images alone, the system achieves accurate prediction while avoiding the complexity of 3D image processing workflows.
Solution Approach 2:
The patent creates a virtual model or representation of wheel performance directly from 2D images through machine learning, rather than creating physical or detailed 3D copies. This virtual modeling approach maintains prediction accuracy while significantly reducing processing complexity compared to full 3D reconstruction.
2Measurement precision
If 2D wheel images are converted to 3D images to predict wheel performance, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent removes the time-consuming 3D conversion step from the workflow by directly extracting performance prediction features from 2D images. This extraction approach maintains the essential information needed for accurate prediction while eliminating the temporal overhead of 3D processing.
Solution Approach 2:
The patent skips the intermediate 3D conversion step entirely, moving directly from 2D image input to performance prediction through machine learning models. This skipping of unnecessary intermediate processing stages significantly reduces overall processing time while preserving prediction accuracy.
3Reliability
If existing technology is used to predict wheel performance, then established methods are maintained, but usefulness is degraded due to 3D conversion requirement
Solution Approach 1:
The patent extracts the core functionality of performance prediction from the existing 3D-based workflow and isolates it to work directly with 2D images. By separating the essential prediction capability from the unnecessary 3D conversion requirement, the system maintains reliability while dramatically improving ease of operation and practical usefulness.
Solution Approach 2:
Instead of following the traditional approach of converting 2D to 3D for analysis, the patent inverts the workflow by directly analyzing 2D images to predict performance. This inversion of the conventional process eliminates the degrading 3D conversion step while maintaining the reliability of performance prediction through alternative machine learning approaches.
Data Source
AI summary
An apparatus for predicting performance of a wheel in a vehicle: includes a learning device that generates a latent space for a plurality of two-dimensional (2D) wheel images based on a convolutional autoencoder (CAE), extracts a predetermined number of the plurality of 2D wheel images from the latent space, and learns a dataset having the plurality of 2D wheel images and performance values corresponding to the plurality of 2D wheel images; and a controller that predicts performance for the plurality of 2D wheel images based on a performance prediction model obtained by the learning device.