Tomographic Image Processing Device for Ophthalmic Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The burden on examiners to accurately diagnose a subject eye from a large number of tomographic images is excessive, as they need to carefully observe each image to identify abnormal portions, which can lead to oversight of minor lesions.
Innovation Solution
A tomographic image processing device using machine learning to extract feature amounts from normal and abnormal eye images, determining the presence of abnormal portions, and displaying marked images to facilitate diagnosis, reducing the examiner's workload and improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If examiners manually observe all tomographic images to diagnose the subject eye, then diagnostic accuracy can be maintained, but the burden on the examiner becomes excessively large
Solution Approach 1:
The patent introduces an automated image processing device as an intermediary between the tomographic images and the examiner. This device automatically analyzes images, detects abnormal portions, and generates summary reports, thereby reducing the examiner's burden while maintaining diagnostic accuracy through automated preprocessing and filtering of images.
Solution Approach 2:
The system enables self-service by allowing the image processing device to autonomously perform image analysis, abnormal portion detection, and report generation without requiring manual examination of each image. The examiner only needs to review the automatically generated results, significantly reducing their workload while maintaining diagnostic reliability.
2Measurement precision
If examiners carefully observe each tomographic image to identify abnormal portions, then detection accuracy improves, but the time required for diagnosis increases
Solution Approach 1:
The system performs preliminary analysis by automatically processing all tomographic images before the examiner reviews them. The image processing device pre-identifies abnormal portions, generates summary reports, and prepares filtered image sets, so that when the examiner reviews the images, the time-consuming analysis has already been completed, reducing overall diagnosis time while maintaining detection accuracy.
Solution Approach 2:
The system creates processed copies of the original tomographic images that highlight abnormal portions and organize images by significance. These processed copies serve as time-efficient substitutes for manual examination of all original images, allowing examiners to quickly review key findings without sacrificing detection accuracy.
3Reliability
If the number of tomographic images captured is increased to improve comprehensive coverage, then diagnostic completeness improves, but the complexity of image review increases
Solution Approach 1:
The system segments the large volume of tomographic images into organized groups based on detected abnormalities, significance levels, and anatomical regions. The image processing device automatically categorizes images and presents them in a structured manner to the examiner, reducing review complexity while maintaining comprehensive diagnostic coverage by ensuring all segments are systematically evaluated.
Solution Approach 2:
The system applies different processing and presentation qualities to different portions of the image set based on their diagnostic importance. Images containing abnormal portions are highlighted with enhanced visualization and prioritized in the review sequence, while normal images are grouped separately. This local differentiation reduces overall review complexity by directing the examiner's attention to critical areas first.
Data Source
AI summary
A tomographic image processing device that includes an input unit configured to input a tomographic image of a subject eye; a processor; and a memory storing computer-readable instructions therein. The computer-readable instructions, when executed by the processor, may cause the processor to execute: acquiring a tomographic image of a normal eye; acquiring a tomographic image of an eye having an abnormal portion; extracting a feature amount of the abnormal portion by using machine learning from the tomographic image of the normal eye and the tomographic image of the eye having the abnormal portion; acquiring a tomographic image of the subject eye inputted to the input unit; and determining whether the tomographic image of the subject eye includes an abnormal portion based on the feature amount.