Reliability-Based 3D Object Model Reduction From Single-View Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing techniques face challenges in efficiently processing model information, including shape and texture information, for object detection, particularly in generating three-dimensional representations.

Innovation Solution

A detection device that includes a detector, reliability calculator, information calculator, and information processor to generate reduced model information by assessing reliability and data points, using methods like lossy compression and reliability-based data reduction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If complete model information including shape and texture data is processed, then detection accuracy is improved, but data processing complexity and time increase

Engineering Contradiction:
Improvedetection accuracyVSAvoiddata processing efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments model information into multiple types (shape information, texture information, and other attributes) and processes them differently based on their importance and characteristics. This allows critical shape data to be processed with high precision while less critical texture data can be processed more efficiently, resolving the contradiction between detection accuracy and processing efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different processing qualities to different parts of the model information. High-precision processing is applied to shape information which is critical for detection accuracy, while lower-precision or compressed processing is applied to texture information. This local differentiation maintains detection accuracy for essential features while improving overall processing efficiency.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If detailed model information is generated, then object representation accuracy is improved, but data amount increases

Engineering Contradiction:
Improveobject representation accuracyVSAvoiddata amount
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent extracts and separates essential model information (shape data) from less essential information (texture data). By taking out only the critical shape information that is necessary for accurate object representation, the system achieves high detection accuracy while significantly reducing the total data amount that needs to be processed and stored.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial processing to model information, focusing computational resources on generating detailed shape information while using compressed or reduced-resolution texture information. This partial action approach achieves sufficient object representation accuracy without the excessive data amounts that would result from processing all model information at full detail.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If model information is compressed to reduce data amount, then processing efficiency is improved, but information loss occurs

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidmodel information accuracy
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent applies different compression qualities to different types of model information. Shape information, which is critical for object detection accuracy, is processed with minimal or no compression to preserve information. Texture information and other less critical attributes are compressed more aggressively. This local differentiation in compression quality improves processing efficiency while minimizing information loss in critical areas.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent introduces an intermediary processing stage that selectively compresses different components of model information based on their importance. This intermediary layer acts as a mediator between the original detailed model data and the final processed output, allowing efficient compression of non-critical data while preserving the integrity of essential shape information.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3428877B1Detection device, information processing device, detection method, detection program, and detection system
Publication Date: 2025.11.05 NIKON CORP
  • EP3428877B1 patent drawingFigure 1(A)~1(B)
  • EP3428877B1 patent drawingFigure 2(A)~2(B)
  • EP3428877B1 patent drawingFigure 3

AI summary

To easily process model information including at least one of shape information and texture information, on an object, or information used for generation of the model information. A detection device (1) includes: a detector (2) that detects an object (OB) from one viewpoint (Vp); an reliability calculator (3) that calculates reliability information on the object from the one viewpoint; an information calculator (4) that uses detection results of the detector to calculate model information including at least one of shape information and texture information, on the object, from the one viewpoint; and an information processor (5) that uses first object information including the model information or source data used for generation of the model information, and the reliability information, so as to generate second object information having a reduced data amount of the model information or the source data.