3D Mechanical Part Querying Using Trihedral Axis Systems

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for querying databases of 3D modeled objects representing mechanical parts lack efficiency in relevance and speed, particularly in identifying similar objects due to inefficiencies in shape descriptor technologies used for mechanical parts.

Innovation Solution

A computer-implemented method that uses a shape descriptor based on trihedral axis systems, which are bijectively associated with vertices and faces of 3D meshes, to evaluate similarity transformations and match axis systems, improving the relevance and speed of querying 3D modeled objects in a database.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing shape descriptor technologies are used for querying 3D modeled objects, then the querying function can be provided, but the relevance and speed of identifying similar mechanical parts are insufficient

Engineering Contradiction:
Improvequerying speedVSAvoidsimilarity detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The shape descriptor is segmented into multiple components including radial function coordinates in a harmonic basis, curvature information, and topology descriptors. This segmentation allows each component to be optimized independently for both computational efficiency and accuracy in detecting similar mechanical parts.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms the shape representation by changing parameters from raw geometric data to spectral domain representations using harmonic functions. This parameter transformation enables faster comparison operations while preserving the essential features needed for accurate similarity detection of mechanical parts.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If detailed shape descriptors are computed for accurate similarity detection, then the relevance of results improves, but the computational time and memory usage increase

Engineering Contradiction:
Improvesimilarity detection accuracyVSAvoidcomputational time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the essential features needed for mechanical part similarity detection from complete shape descriptors. By taking out and retaining only the most discriminative components (radial functions, curvature, topology), the system achieves accurate similarity detection with reduced computational overhead.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Shape descriptors are pre-computed and stored in an optimized spectral representation before querying. This preliminary action transforms raw geometric data into a compact form that enables fast comparison operations during actual queries, significantly reducing real-time computational time while maintaining detection accuracy.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If complete shape descriptors are stored for each 3D object, then accurate similarity search is enabled, but the data size and storage requirements become large

Engineering Contradiction:
Improveshape comparison accuracyVSAvoiddata size
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

Instead of storing complete high-resolution geometric models for similarity comparison, the patent stores compressed spectral representations (copies) of shape descriptors. These copied representations in the harmonic basis capture essential shape characteristics while occupying significantly less storage space, enabling accurate similarity search without proportionally large data sizes.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS10303784B2Querying a database with likeness criterion
Publication Date: 2019.05.28 DASSAULT SYSTEMES SA
  • US10303784B2 patent drawing
  • US10303784B2 patent drawing
  • US10303784B2 patent drawing

AI summary

A computer-implemented method and system for querying a database that comprises 3D modeled objects that represent mechanical parts. The method or system provides a query that includes a likeness criterion relative to the value of a predetermined shape descriptor for an input 3D modeled object. A next step returns, as results of the query, respective 3D modeled objects of the database based on an extent to which the value of the predetermined shape descriptor for the respective 3D modeled object respects the likeness criterion. The value of the predetermined shape descriptor for a respective 3D modeled object includes a set of axis systems that are trihedral. The likeness increasingly depends on an extent to which sets of axis systems match each other modulo a same similarity transformation. Such a method and system improves the querying of a database that comprises 3D modeled objects that represent mechanical parts.