3D Model Creation Refining Local Features for Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing 3D model creation technologies face challenges in accurately reflecting local area information that affects recognition performance, leading to insufficient recognition performance and requiring large data and processing volumes due to unwanted noise reflection.

Innovation Solution

A model creation apparatus that uses a processor and memory to acquire and correct 3D models based on feature information from registration target objects, refining the shape models by adjusting vertices and meshes to match local area features, thereby reflecting local information effectively with reduced data and processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If a standard three-dimensional shape model is modified based on the difference between the photographic object image and the target image, then the shape model can be reconstructed, but the recognition performance may be insufficient due to insufficient evaluation of the local area

Engineering Contradiction:
Improveshape model reconstruction accuracyVSAvoidrecognition performance
Core Design Contradiction:
Manufacturing precisionVSReliability

Solution Approach 1:

The patent applies local quality by evaluating and modifying only specific local areas of the three-dimensional shape model that have significant impact on recognition performance. The system calculates local area evaluation values for different regions and selectively applies modifications based on these evaluations, rather than uniformly processing the entire model. This ensures that critical local features are accurately reconstructed while maintaining overall model quality.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent changes parameters by introducing local area evaluation values that quantify the importance of different model regions. By modifying the model based on these parameter-based evaluations rather than simple image differences, the system achieves better recognition performance. The parameter changes enable selective refinement of model regions that matter most for recognition tasks.

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If the standard three-dimensional shape model is modified based on image differences, then the shape model can be updated, but a large volume of data and processing is required because fluctuations (noise) of local areas are undesirably reflected on the new 3D model

Engineering Contradiction:
Improvelocal area reflection accuracyVSAvoiddata and processing volume
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent applies local quality by evaluating and modifying only specific local areas of the three-dimensional shape model that have significant impact on recognition performance. The system calculates local area evaluation values for different regions and selectively applies modifications based on these evaluations, rather than uniformly processing the entire model. This ensures that critical local features are accurately reconstructed while maintaining overall model quality.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent extracts only the essential local area information that significantly affects recognition performance, separating it from noise and irrelevant fluctuations. By calculating local area evaluation values and using these to guide selective modification, the system extracts meaningful features while discarding noisy data, thereby reducing the volume of data and processing required.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS12154294B2Model creation device and model creation method
Publication Date: 2024.11.26 HITACHI LTD
  • US12154294B2 patent drawing
  • US12154294B2 patent drawing
  • US12154294B2 patent drawing

AI summary

A model creation apparatus being configured to: hold at least one image of the registration target object in one or more postures and a reference model indicating a shape of a reference object; acquire information indicating a feature of the registration target object in a first posture; and correct, when a shape in the first posture that is indicated by the reference model is determined to be dissimilar based on a predetermined first condition, the reference model based on the information indicating the feature to thereby create the model indicating the shape of the registration target object.