Endoscope Image Review With Correctable Observation Completion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing medical image processing systems face challenges in accurately determining the completion of observation of specific parts in medical images, leading to potential oversight due to discrepancies between artificial intelligence and human judgment, which can result in incomplete or insufficient examination.

Innovation Solution

A medical image processing apparatus that allows for user input to correct observation completion determinations, reflecting the corrected content in the display, and utilizes criteria such as the presence of a mark, duration, and central position of the specific part in the image to determine observation completion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI-based image recognition is used to automatically determine observation completion, then productivity is improved, but reliability deteriorates due to discrepancies between AI and human judgment

Engineering Contradiction:
Improveobservation completion determination efficiencyVSAvoiddetermination accuracy
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system displays the AI's observation completion determination results to the user and accepts corrections. When the user corrects an erroneous determination, the corrected information is fed back to refine the AI's judgment, resolving the reliability issue while maintaining automated efficiency

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The display device acts as an intermediary between the AI determination system and the user. It presents AI results for review and allows user correction, bridging the gap between automated processing and human judgment without eliminating either

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If AI determination criteria are used, then ease of operation is improved, but measurement precision deteriorates due to subjective factors in human observation

Engineering Contradiction:
Improveautomatic determination capabilityVSAvoidobservation completion accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The determination criteria are made dynamic by allowing the system to switch between AI-based automatic determination and user-based manual determination. The system adapts to different observation scenarios and corrects discrepancies in real-time, improving precision without sacrificing operational ease

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system performs AI determination for all parts automatically, then allows selective user correction only where needed. This partial human intervention corrects precision errors without requiring full manual review, maintaining ease of operation

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4091532B1Medical image processing device, endoscope system, diagnosis assistance method, and program
Publication Date: 2025.12.24 FUJIFILM CORP
  • EP4091532B1 patent drawingFigure 1
  • EP4091532B1 patent drawingFigure 2
  • EP4091532B1 patent drawingFigure 3

AI summary

There are provided a medical image processing apparatus, an endoscope system, a diagnosis assistance method, and a program capable of coping with an erroneous determination while utilizing an observation completion determination in which image processing is used and of suppressing oversight of an observation target part by a doctor. The medical image processing apparatus includes at least one processor. The at least one processor acquires a medical image; makes, on the basis of the medical image, an observation completion determination as to whether observation is completed for a target part; performs display control for causing a determined result of the observation completion determination to be displayed on a display device, receives a user input including an instruction for correcting display content indicating the determined result of the observation completion determination, and causes corrected content based on the user input to be reflected in the display.