Annular Face Estimation via Multi-Model Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for segmenting cross-sectional images of cylindrical shapes, such as those using circle or ellipse fitting and machine learning, often result in errors due to shape limitations and interruptions, leading to inaccuracies in estimating the thickness and presence of holes in annular faces.

Innovation Solution

An estimation processing program and apparatus that utilize multiple neural networks to segment annular, outer, and inner circumferences of cylindrical shapes, combining their outputs to improve accuracy and reduce errors through noise removal and arithmetic operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If circle or ellipse fitting method is used for segmentation, then the segmentation process is simple, but the shape of cross section is limited to circle or ellipse and inner and outer circles may cross

Engineering Contradiction:
ImproveSegmentation process simplicityVSAvoidShape adaptability
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent segments the segmentation task into multiple independent neural networks, each responsible for specific boundaries (outer circumference, inner circumference). This allows the system to handle complex annular shapes while maintaining manageable processing through modular architecture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The neural network-based segmentation system provides universal applicability to various cross-sectional shapes including circles, ellipses, and irregular annular shapes, replacing the shape-limited fitting method with a versatile deep learning approach.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Adaptability or versatility

If machine learning method is used for edging learning, then segmentation can be performed, but edging region is liable to be interrupted and cylindrical shape prior knowledge cannot be utilized

Engineering Contradiction:
ImproveSegmentation capabilityVSAvoidEdging continuity
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies different processing strategies to different regions: the outer circumference network handles the outer boundary while the inner circumference network handles the inner boundary. This localized approach ensures continuous edging in each region while maintaining overall segmentation accuracy.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system performs preliminary segmentation using multiple specialized neural networks before final combination. Each network is trained specifically for its designated boundary, ensuring continuous and accurate edging detection while incorporating cylindrical shape constraints through the multi-network architecture.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If single network segmentation is used, then processing is efficient, but error decision occurs about thickness and hole presence

Engineering Contradiction:
ImproveProcessing efficiencyVSAvoidThickness and hole detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides the segmentation task into separate neural networks for outer circumference and inner circumference detection. This segmentation allows each network to specialize in specific boundary detection, improving measurement precision for thickness and hole presence while maintaining processing efficiency through parallel operation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an arithmetic operation unit as an intermediary that combines the outputs of multiple neural networks. This intermediary component synthesizes the segmented results to produce accurate thickness measurements and hole presence detection, resolving the contradiction between efficient processing and precise measurement.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3859668B1Estimation processing program, estimation processing method and information processing apparatus
Publication Date: 2024.06.19 FUJITSU LTD
  • EP3859668B1 patent drawingFigure 1
  • EP3859668B1 patent drawingFigure 2
  • EP3859668B1 patent drawingFigure 3

AI summary

A method includes: inputting an image of a target to each of a first model for estimating a region of an annular face of the target, a second model for estimating a region in an outer circumference of the annular face and a third model for estimating a region in an inner circumference of the annular face, and outputting a region image, as an estimated region of the annular face, the region image being obtained by synthesizing first, second, and third elements. The first element is a region outputted as a result of the first model for calculating a sum set in the synthesizing, the second element is a region outputted as a result of the second model for calculating a sum set in the synthesizing, and the third element is a region outputted as a result of the third model for calculating a difference set in the synthesizing.