Endoscopic Image Recognition for Captured Region Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Capturing desired medical images with an endoscope system is challenging, especially for inexperienced users, as it requires adherence to specific imaging conditions and determining if the captured images meet predefined criteria, which is difficult without proper guidance.
Innovation Solution
A medical image processing apparatus and method that includes a processor for determining the presence of multiple imaging target sites in a medical image and assessing if the image meets specific criteria, displaying a notification indicator when the criteria are met, using a combination of convolutional neural networks and multiple indices for evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual imaging according to predefined conditions is performed, then imaging quality can be controlled, but it becomes difficult for inexperienced users to capture desired images and determine if appropriate images are captured
Solution Approach 1:
The system automatically evaluates captured images against predefined criteria and provides notifications when criteria are met, enabling the system to self-assess imaging quality without requiring user judgment. The processor automatically determines whether imaging target sites are included and whether determination criteria are satisfied, freeing users from manual evaluation tasks.
Solution Approach 2:
The system provides real-time feedback to users through notification indicators that inform them when captured images satisfy the determination criteria. This feedback mechanism guides users in capturing appropriate images by immediately indicating when the imaging requirements are met, reducing the difficulty of determining image adequacy.
2Reliability
If multiple imaging target sites are evaluated using multiple indices, then image quality assurance is improved, but the complexity of the evaluation system increases
Solution Approach 1:
The evaluation system is segmented into multiple independent determiners, each responsible for evaluating specific indices (e.g., focus, brightness, composition) for specific imaging target sites. This segmentation allows complex multi-criteria evaluation to be broken down into manageable, parallel processing tasks, reducing overall system complexity while maintaining comprehensive quality assurance.
Solution Approach 2:
The evaluation system uses a universal processor that can assess multiple imaging target sites using multiple indices through a single integrated framework. The same processing architecture evaluates different sites (esophagogastric junction, cardia, pylorus, etc.) and different criteria (focus, brightness, composition) without requiring separate specialized systems for each evaluation type.
Data Source
AI summary
A medical image processing apparatus includes a processor configured to acquire a medical image of a subject, perform image recognition on the medical image using a trained model, based on a result of the image recognition, in response to a region included in the medical image satisfying a determination criterion, identify the region included in the medical image as a captured region, and output information indicating that the region is the captured region with respect to a schematic diagram representing an organ.