Endoscopic Image Processing With Integrated Lesion Inference Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Endoscopic images often include diverse lesions and varying photographing environments, making accurate detection of lesion regions challenging and inconsistent among doctors.
Innovation Solution
An image processing device that acquires endoscopic images, generates multiple inference results using data augmentation and integration of lesion region inference models, and integrates these results to accurately detect lesion regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple inference models are used to detect lesion regions, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The system divides the lesion detection task into multiple specialized inference models, each trained to detect specific types of lesions or operate under specific conditions. This segmentation allows each model to focus on a particular aspect, improving overall detection accuracy while maintaining manageable complexity through modular architecture
Solution Approach 2:
Multiple inference results from different models are merged through integration processing to produce a final detection result. This combining approach leverages the strengths of each individual model, achieving higher accuracy than any single model could provide alone
2Measurement precision
If data augmentation is applied to improve detection reliability, then measurement precision increases, but processing time increases
Solution Approach 1:
Data augmentation is performed in advance to create diverse training datasets and pre-train multiple inference models with varied image conditions. This preliminary preparation ensures the models are already adapted to different scenarios, reducing the need for extensive real-time processing while maintaining high detection accuracy
Data Source
AI summary
The image processing device 1X includes an acquisition means 30X, an inference means 32X, and an integration means 33X. The acquisition means 30X acquires an endoscopic image obtained by photographing an examination target. The inference means 32X generates plural inference results regarding an attention region of the examination target in the endoscopic image, based on the endoscopic image. The integration means 33X integrates plural inference results.