Adaptive Lesion Detection in Endoscopic Imaging Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lesion detection methods in endoscopic examinations face challenges: methods based on a fixed number of images are susceptible to noise like blurring but maintain high accuracy, while methods based on a variable number of images are less susceptible to noise but may lead to detection delays or misses when there is no substantial change between images.
Innovation Solution
An image processing device and method that acquire endoscopic images and detect lesions using a selection model chosen from a first model and a second model. The first model makes inferences based on a predetermined number of images, while the second model does so based on a variable number of images. The device changes parameters for detection based on a non-selection model, allowing for adaptive lesion detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed predetermined number of images is used for lesion detection, then detection accuracy is maintained high, but the system becomes susceptible to noise such as blurring
Solution Approach 1:
The patent implements dynamic switching between two detection models: a first model that uses a fixed predetermined number of images for high accuracy detection, and a second model that uses a variable number of images to reduce noise susceptibility. The system dynamically selects which model to use based on real-time image quality assessment, thereby adapting to changing conditions and resolving the contradiction between maintaining high detection accuracy and reducing noise susceptibility.
2Object-affected harmful factors
If a variable number of images is used for lesion detection, then noise susceptibility is reduced, but detection delay or miss occurs when there is no substantial change between images
Solution Approach 1:
The system dynamically adjusts the detection strategy by switching between a first model using fixed number of images and a second model using variable number of images. When images show substantial changes indicating potential lesions, the system uses the second model for rapid detection. When images remain relatively stable, the system switches to the first model to ensure timely detection, thereby resolving the contradiction between reducing noise susceptibility and preventing detection delay.
3Device complexity
If a single detection model is used, then the system structure remains simple, but the system cannot adapt to different image conditions optimally
Solution Approach 1:
The patent implements a universal lesion detection system that incorporates two different detection models within a single system. The first model handles cases requiring high detection accuracy with fixed image counts, while the second model handles cases requiring noise reduction with variable image counts. An image quality assessment mechanism determines which model to activate, enabling the system to adapt to different image conditions optimally while maintaining a unified architecture.
Data Source
AI summary
The image processing device 1X includes an acquisition means 30X and a lesion detection means 34X. The acquisition means 30X acquires an endoscopic image obtained by photographing an examination target by a photographing unit provided in an endoscope. The lesion detection means 34X detects a lesion based on a selection model which is selected from a first model and a second model, the first model being configured to make an inference regarding a lesion of the examination target based on a predetermined number of endoscopic images, the second model being configured to make an inference regarding a lesion of the examination target based on a variable number of endoscopic images. Besides, the lesion detection means 34X changes a parameter to be used for detection of the lesion based on a non-selection model that is the first model or the second model other than the selection model.