Image Processing for Endoscopic Swallowing Stage Assessment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video endoscopic examinations for swallowing function evaluation rely heavily on subjective doctor evaluations, leading to inconsistent and less accurate diagnoses due to variability in physician skill.

Innovation Solution

An image processing device that analyzes video endoscopic examination footage frame by frame, assigning swallowing stage information and calculating feature values based on machine learning algorithms to provide objective and accurate indicators of swallowing function.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If subjective evaluation by doctor's visual examination is used, then the diagnosis can be performed, but the diagnosis accuracy is low and varies depending on doctor's skill

Engineering Contradiction:
Improvediagnosis accuracyVSAvoiddiagnosis consistency
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent replaces the mechanical/subjective visual examination system with an automated image processing system. The control processor automatically analyzes endoscopic images using recognition processing to identify swallowing stages and calculate feature values, eliminating dependence on doctor's visual skills and ensuring consistent, objective diagnosis accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service diagnosis by automatically performing the entire evaluation process. The control processor independently performs image acquisition, recognition processing, stage information assignment, and feature value calculation without requiring doctor's manual analysis, allowing the system to evaluate itself objectively and consistently.

Inventive Principle:
Principle #25Self-service

2Measurement precision

If automatic image processing is introduced, then diagnosis accuracy and objectivity are improved, but device complexity increases

Engineering Contradiction:
Improvediagnosis accuracyVSAvoidprocessing system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The control processor is designed as a multi-functional device that integrates image acquisition, recognition processing, stage information assignment, and feature value calculation capabilities. By consolidating these functions into a single processor, the patent reduces overall system complexity while maintaining high diagnosis accuracy and objectivity.

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

Solution Approach 2:

The patent segments the swallowing process into distinct stages (swallowing-in-progress stage, swallowing-not-in-progress stage, etc.) and processes images frame-by-frame according to these segments. This segmentation approach simplifies the complex recognition task into manageable steps, reducing processing complexity while improving accuracy.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If frame-by-frame analysis is performed, then measurement precision of swallowing stages is improved, but processing time increases

Engineering Contradiction:
Improveswallowing stage identification accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by pre-defining multiple swallowing stages and their corresponding characteristics before actual analysis. The control processor has pre-established recognition criteria for each stage, allowing rapid frame-by-frame analysis without time-consuming real-time decision-making, thus reducing processing time while maintaining precision.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The control processor continuously processes images in a seamless frame-by-frame manner, maintaining constant recognition and stage assignment without interruption. This continuous processing approach ensures that all frames are analyzed thoroughly for high precision while the automated nature of continuous processing minimizes overall time loss compared to manual evaluation.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP4585134A1Image processing device, operation method of image processing device, non-transitory computer readable medium, and diagnosis support apparatus
Publication Date: 2025.07.16 FUJIFILM CORP
  • EP4585134A1 patent drawingFigure 1(A)~1(B)
  • EP4585134A1 patent drawingFigure 2
  • EP4585134A1 patent drawingFigure 3A~3C

AI summary

Provided are an image processing device, an operation method of an image processing device, an image processing program, and a diagnosis support apparatus capable of easily obtaining an objective and highly accurate indicator related to swallowing without taking time and effort. An image processing device acquires an examination video in which an observation target is imaged in a video endoscopic examination of swallowing, performs recognition processing on the examination video for each frame image, to assign stage information indicating that an image shows which stage of the observation target among a plurality of swallowing stages set in advance to each of the frame images, and calculates a feature value indicating a feature related to swallowing of a subj ect person having the observation target based on the stage information assigned to each of a plurality of the frame images included in the examination video.