Endoscopic Image Processing for Objective Swallowing Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video endoscopic examinations for swallowing function evaluation rely heavily on subjective doctor evaluations, leading to variability in diagnosis accuracy due to the skill of the doctor.
Innovation Solution
An image processing device that analyzes video endoscopic examination footage frame by frame, assigning stage information to each image and calculating feature values related to swallowing, providing objective and accurate indicators of swallowing function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subjective evaluation by doctor's visual examination is used, then the examination can be performed with simple equipment, but the diagnosis accuracy varies depending on the doctor's skill
Solution Approach 1:
The patent replaces the mechanical visual examination system with an automated image processing system that uses computer vision algorithms. The system automatically detects swallowing stages and calculates feature values from endoscopic images, eliminating the need for manual evaluation by doctors and thereby improving diagnosis accuracy consistency.
Solution Approach 2:
The patent creates a digital copy of the swallowing process through automated image analysis. By capturing multiple frame images and processing them through machine learning models, the system creates a comprehensive digital representation of swallowing function that can be objectively measured and compared, replacing the subjective human evaluation.
2Measurement precision
If automated image processing is introduced, then objective and accurate swallowing indicators can be obtained, but the system complexity increases
Solution Approach 1:
The patent segments the swallowing process into distinct stages (pre-swallowing, swallowing-in-progress, post-swallowing) and processes each stage separately through automated image analysis. This segmentation allows complex swallowing motion to be broken down into manageable components that can be systematically evaluated, improving measurement accuracy while keeping the processing logic organized.
Solution Approach 2:
The patent transforms qualitative visual evaluation into quantitative parameters by calculating feature values from image data. The system extracts measurable parameters such as the number of frame images in each swallowing stage and uses these as objective indicators, converting the complex visual assessment into measurable, comparable data.
3Measurement precision
If frame-by-frame analysis is performed, then highly accurate swallowing indicators can be calculated, but the processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-defining swallowing stages and creating machine learning models that are trained in advance to recognize these stages. The system uses pre-processed image data and pre-established classification criteria, which allows for rapid automated analysis without requiring time-consuming real-time decision-making during the examination.
Data Source
AI summary
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 subject 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.