Medical Image Region Analysis With User-Guided Refinement

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing medical image analysis technologies lack accuracy in analyzing regions of interest, necessitating improvements in analysis results.

Innovation Solution

A medical image analysis apparatus and method that performs a first analysis process on multiple regions of interest, allows user input for selecting a specific region, and executes a second analysis process based on this input, using trained models to refine and correct analysis results, including changing analysis parameters and incorporating user comments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a single analysis process is performed on multiple regions of interest, then the analysis speed is maintained, but the accuracy of analysis results for specific regions deteriorates

Engineering Contradiction:

Improveaccuracy of analysis resultsVSAvoidcomplexity of analysis process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The analysis process is divided into a first analysis process that processes multiple regions of interest simultaneously, and a second analysis process that performs detailed analysis on a specific region selected by the user. This segmentation allows the system to maintain both efficiency in processing multiple regions and accuracy in analyzing the selected region of interest.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The first analysis process is performed as a preliminary step to identify multiple regions of interest and generate initial analysis results. Based on these results, the user can then select a specific region for the second, more detailed analysis process. This preliminary action enables efficient routing of computational resources to where they are most needed.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If a second analysis process is performed on a selected region of interest, then the accuracy of analysis results is improved, but the analysis time increases

Engineering Contradiction:

Improveaccuracy of analysis resultsVSAvoidanalysis time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Instead of performing a comprehensive detailed analysis on all regions of interest, the system applies the second, more time-consuming analysis process only to the specific region selected by the user. This partial action approach maintains high accuracy for the selected region while minimizing the overall time loss by avoiding redundant detailed analysis of other regions.

Inventive Principle:

Principle #16Partial or excessive action

3Loss of information

If multiple analysis results are generated for the first region of interest, then the completeness of information is improved, but the difficulty of selecting the correct result increases

Engineering Contradiction:

Improvecompleteness of analysis informationVSAvoidease of selecting analysis result

Core Design Contradiction:

Loss of informationVSEase of operation

Solution Approach 1:

The system presents multiple analysis results from the first analysis process to the user and incorporates user feedback through selection input. The user can review the multiple results and select the most appropriate one, with the system then using this feedback to guide the second analysis process. This feedback mechanism ensures both completeness of information and ease of operation.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12536659B2Medical image analysis apparatus, medical image analysis method, and medical image analysis program

Publication Date: 2026.01.27 FUJIFILM CORP

AI summary

A medical image analysis apparatus extracts a region having one or more preset physical features from a medical image, performs control to display information indicating the extracted region, receives information indicating a selected region from among the extracted regions, refers to data in which physical features are associated with disease names, and generates a comment on findings using the disease name associated with the physical features of the selected region.