Capsule Endoscopy Image Interpretation System with Abnormality Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of image interpretation using capsule endoscopes is time-consuming and requires significant concentration due to the large number of images captured, which can be as many as 60,000, and existing software lacks efficient tools to streamline this process.
Innovation Solution
An image display apparatus and system that includes an image playback unit, operating unit, image interpretation operation recording unit, and image interpretation operation image display unit, which allows for chronological playback, operation recording, and display of images, along with features like playback speed regulation, automatic extraction of red-dominated images, and the ability to cut out and transmit parts of image groups for external interpretation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual image interpretation is performed on all captured images, then interpretation accuracy is maintained, but interpretation time increases significantly
Solution Approach 1:
The system extracts and displays only images that contain abnormalities (red-dominated images) from the large set of captured images, separating the important diagnostic information from the routine normal images. This allows interpreters to focus on critical cases while automated processing handles routine images.
Solution Approach 2:
The interpretation process is segmented into automated preprocessing (filtering, abnormality detection) and manual review (focused interpretation of flagged images). This division allows different levels of human involvement based on image characteristics.
2Reliability
If all 60,000 captured images are reviewed manually, then complete diagnostic coverage is achieved, but interpreter concentration and workload increase
Solution Approach 1:
The system performs automated preprocessing, filtering, and abnormality detection without continuous human intervention. The automated extraction function identifies red-dominated images and prepares them for review, reducing the burden on interpreters.
Solution Approach 2:
The system provides visual feedback through the abnormality indication image that shows which images contain abnormalities and their locations. This feedback mechanism guides interpreters efficiently through the image set without requiring manual inspection of every image.
3Loss of information
If detailed operation recording is implemented, then interpretation process tracking is improved, but system complexity increases
Solution Approach 1:
The system creates a simplified visual copy of the interpretation process through the abnormality indication image, which reproduces the essential information about which images were flagged and why, without requiring complex detailed logging of every interpreter action.
Data Source
AI summary
An image display apparatus includes an image playback unit that plays a group of a series of images in chronological order; an operating unit that is used to input an operation when an image that is being played in the image playback unit is subjected to image interpretation; an image interpretation operation recording unit that records an operation content which is input from the operating unit in an order with respect to the image that is being played in the image playback unit; an image interpretation operation image creating unit that creates, using the operation content recorded in the image interpretation operation recording unit, a series of images according to the order; and an image interpretation operation image display unit that displays the image created by the image interpretation operation image creating unit.