Capsule Endoscopy Polyp Detection for Faster Image Review
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing capsule endoscopy systems require extensive manual review of thousands of images by healthcare professionals, leading to tiresome and time-consuming report generation, with potential errors in identifying polyps and recommending colonoscopies.
Innovation Solution
A system and method utilizing deep learning neural networks and machine learning algorithms to automatically identify polyps with high confidence, allowing for automated image presentation and overriding incorrect designations, thereby reducing the need for manual review and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual review of thousands of images is performed by healthcare professionals, then polyp identification can be conducted with human expertise, but the process becomes tiresome and time-consuming
Solution Approach 1:
An automated polyp detection system acts as an intermediary between the capsule endoscopy images and the healthcare professional's final diagnosis. The system pre-processes thousands of images, identifies potential polyps, and presents only relevant cases to reviewers, thereby reducing review time while maintaining diagnostic accuracy through human-in-the-loop verification
Solution Approach 2:
The image review process is segmented into multiple stages: automated pre-screening by AI algorithms, filtering of low-probability cases, and focused human review of high-confidence polyp detections. This segmentation allows the system to handle thousands of images efficiently by dividing the task between automated processing and selective human expertise
2Reliability
If manual review of thousands of images is performed by healthcare professionals, then polyp identification can be conducted, but the reading task becomes tiresome
Solution Approach 1:
The system extracts and removes low-value review tasks from the healthcare professional's workflow by automatically filtering out images with high confidence of being normal or containing artifacts. Only images requiring human expertise are presented to reviewers, significantly reducing workload while preserving the need for human judgment in ambiguous cases
Solution Approach 2:
The system performs self-service by automatically conducting initial image analysis, quality assessment, and polyp detection. This automates routine tasks that would otherwise require human effort, allowing healthcare professionals to focus exclusively on complex cases that demand clinical expertise
3Productivity
If automated polyp detection is implemented, then review time can be reduced, but potential errors in identifying polyps and recommending colonoscopies may occur
Solution Approach 1:
The system implements feedback loops where automated detection results are reviewed and validated by healthcare professionals, and their corrections are fed back to improve the algorithm's performance. This continuous feedback mechanism allows the system to learn from errors and improve accuracy over time while maintaining high productivity
Solution Approach 2:
The system dynamically adjusts its confidence thresholds and review requirements based on the detected case characteristics. High-confidence automated detections may require minimal human review, while uncertain cases trigger more rigorous verification. This dynamic approach optimizes both speed and accuracy based on real-time assessment of detection reliability
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods are disclosed for identifying images that contain polyps. An exemplary method for identifying images includes: accessing images of a gastrointestinal tract (GIT) captured by a capsule endoscopy device, where: each image of the images is suspected to include a polyp and is associated with a probability of containing the polyp, and the images include seed images, where each seed image is associated with one or more images of the images. The image(s) associated with each seed image is identified as suspected to include the same polyp as the associated seed image. The method includes applying a polyp detection system on the seed images to identify seed images which include polyps, where the polyp detection system is applied to each seed image of based on the image(s) associated with the seed image and the probabilities associated with the seed image and with the associated image(s).