AI-Guided Multi-Camera Endoscopy for Wide-View Polyp Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current endoscopes with a single standard camera struggle to provide a comprehensive view of the colon, particularly in areas like the folds, requiring multiple cameras that overwhelm physicians with complex and distorted images, making it difficult to identify polyps accurately.
Innovation Solution
A multi-camera endoscope system with AI-powered target feature detection, where secondary cameras capture additional views that are analyzed by a neural network to detect polyps, generating alerts to guide the physician's attention to potential polyp locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple cameras are used to increase field of view, then the ability to visualize the entire colon is improved, but the device complexity and difficulty of operation increase as physicians must look at multiple images at once
Solution Approach 1:
An AI-based processing system acts as an intermediary between the multiple cameras and the physician. The system receives video feeds from multiple cameras, processes them to identify polyps and target features, and presents only relevant information to the physician through alerts and highlighted regions, eliminating the need to manually monitor multiple camera feeds simultaneously
Solution Approach 2:
The system extracts and isolates only the most critical information (polyps and target features) from the complex multi-camera video feeds. By separating the detection function from the display function, the system presents physicians with only the essential alerts and locations rather than requiring them to process all camera images at once
2Area of stationary object
If wide-angle cameras are used to increase field of view, then the ability to examine the entire colon is improved, but image distortion at edges makes it difficult to accurately identify polyps
Solution Approach 1:
AI-based processing systems serve as intermediaries that receive distorted wide-angle images, correct the distortions, and identify polyps with high accuracy. The system processes images from multiple angles and synthesizes accurate representations, eliminating the need for physicians to manually correct distortions
Solution Approach 2:
The patent replaces manual image processing and visual inspection with automated AI-based image analysis. Neural networks and computer vision algorithms automatically detect, locate, and characterize polyps in distorted wide-angle images, providing accurate measurements and diagnoses without human intervention in the analysis process
3Area of stationary object
If multiple standard angle cameras are used, then the field of view is increased, but the quantity of information to process increases, overwhelming the physician
Solution Approach 1:
The AI processing system extracts only the most critical information (polyps and target features) from the massive amount of video data generated by multiple cameras. It filters out irrelevant information and presents only essential alerts to the physician, reducing the information quantity from megabytes of continuous video to minimal discrete alerts
Solution Approach 2:
The system performs self-service by automatically processing, analyzing, and prioritizing information from multiple cameras without requiring physician intervention. The AI system autonomously monitors all camera feeds, detects polyps, and generates alerts, freeing the physician from the burden of manually processing large quantities of visual information
Data Source
AI summary
Methods and systems of analyzing and displaying medical videos are described herein. In some embodiments, the methods and systems may be used during an endoscopy procedure using a multi-camera endoscope having a first camera and a second camera. A method according to the present disclosure may include: receiving a first medical video from a first camera; receiving a second medical video from a second camera; detecting one or more target features of the second medical video using a pre-trained target feature detector comprising a neural network; generating an alert indicating the detection of the one or more target features in the second medical video; and displaying the first medical video and the alert on a display.