Fluoroscopy Exposure Control for Low-Dose Image Restoration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fluoroscopy systems face challenges in achieving optimal image quality due to low exposure doses, which complicates downstream image denoising and restoration processes.
Innovation Solution
A method involving a neural network to identify informative frames and adjust exposure parameters dynamically, using a trained model to optimize dose, frame rate, and other settings based on object movement and changes during procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If low exposure dose is used during fluoroscopy procedures, then radiation exposure to patient is reduced, but image quality deteriorates and downstream image denoising/restoration processes are hindered
Solution Approach 1:
The system performs preliminary actions by acquiring a first X-ray image at a first exposure dose level before the fluoroscopy procedure begins. This preliminary image is then used to train neural network models that can optimize subsequent exposure parameters, allowing the system to prepare optimal imaging parameters in advance rather than adjusting them reactively during the procedure.
Solution Approach 2:
The system dynamically adjusts exposure parameters (mA, ms, kV, filters) frame-by-frame during fluoroscopy based on real-time image analysis. The neural network models continuously optimize the exposure dose for each frame based on image content, allowing the system to adapt the radiation dose dynamically to minimize exposure while maintaining optimal image quality for denoising and restoration.
2Ease of operation
If fixed exposure parameters are used throughout fluoroscopy procedure, then system operation is simplified, but image quality cannot be optimized for different frames
Solution Approach 1:
The system performs self-service by automatically analyzing image content and optimizing exposure parameters without manual intervention. The neural network models process each frame's image data and autonomously determine optimal exposure parameters (mA, ms, kV, filters), eliminating the need for operator intervention while achieving frame-specific optimization.
Solution Approach 2:
The system implements feedback by using the first X-ray image and subsequent frame images as input to train neural network models that generate optimized exposure parameters. The output of each frame's image analysis feeds back into the parameter optimization process, creating a closed-loop system that continuously improves exposure settings based on actual image quality and content.
Data Source
AI summary
A method of imaging includes obtaining a sequence of X-ray images including a first X-ray image and a second X-ray image, the first and second X-ray images being acquired sequentially using a first set of exposure parameters of an image scanning apparatus; inputting the first and second X-ray images into a trained first model to obtain a second set of exposure parameters, which is output from the trained first model; obtaining a third X-ray image acquired by the image scanning apparatus using the second set of exposure parameters; inputting at least the second X-ray image and the third X-ray image into a trained second model to obtain a restored X-ray third image, which is output from the trained second model; and outputting the restored X-ray third image, wherein the first model and the second model were trained together using a training sequence of images in a training process.