X-ray Fluoroscopic Imaging Apparatus with Adaptive Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of identifying a target device in an X-ray fluoroscopic imaging apparatus using a trained learning model can deteriorate due to shape differences between the trained device and the actual device used in a surgical operation, leading to erroneous detection and decreased visibility of the target in the X-ray image.
Innovation Solution
The X-ray fluoroscopic imaging apparatus includes an image quality improvement processing unit that switches between two image processing modes: one using a learning identification result for enhancing the target device and another not using the learning identification result for noise reduction, allowing for improved visibility of the target device regardless of the accuracy of the learning identification result.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If a trained learning model is used to identify a target device in an X-ray fluoroscopic image, then the visibility of the target device is improved through enhancement processing, but the identification accuracy deteriorates when there is a shape difference between the trained device and the actual device, leading to erroneous detection
Solution Approach 1:
The patent applies dynamics by making the image processing mode switchable between first mode (using learning identification result) and second mode (not using learning identification result). This dynamic switching allows the system to adapt to different situations: when the device shape matches the training data, the first mode enhances visibility; when there is a shape difference causing erroneous detection, the system switches to the second mode to maintain identification accuracy.
Solution Approach 2:
The patent changes the parameter of image processing mode (first mode vs. second mode) based on the reliability of the learning identification result. When the learned device shape closely matches the actual device, the system uses the first mode for enhanced visibility. When there is a significant shape difference reducing reliability, the system switches to the second mode, effectively changing the processing parameter to prevent erroneous detection while maintaining adequate visibility.
2Illumination intensity
If enhancement processing is performed based on learning identification result to improve target visibility, then the visibility of the target device is improved, but erroneous detection occurs when the learning identification result is inaccurate, causing non-target parts to be emphasized
Solution Approach 1:
The patent implements feedback by evaluating the reliability of the learning identification result (which considers shape matching between trained and actual devices) and using this evaluation to determine the appropriate image processing mode. When the reliability is high, the first mode is used for enhancement; when reliability is low, the system switches to the second mode, preventing erroneous emphasis of non-target parts while maintaining appropriate visibility.
Solution Approach 2:
The system dynamically switches between two image processing modes based on the reliability assessment. This dynamic adjustment ensures that enhancement processing is only applied when the learning identification result is reliable, thereby improving visibility without causing erroneous detection of non-target parts.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This configuration enables the apparatus to maintain or improve the visibility of the target device by adjusting the image processing mode based on the accuracy of the learning identification result, thereby preventing the deterioration of target visibility due to the learning identification result.
Implementation Method 1
an imaging unit (1) including an X-ray source (1a) for irradiating a subject (90) with X-rays
Implementation Method 2
an X-ray detector (1b) for detecting X-rays irradiated from the X-ray source (1a)
Data Source
AI summary
An X-ray fluoroscopic imaging apparatus includes an imaging unit, an X-ray image acquisition unit configured to acquire an X-ray image, a target distribution learning identification unit for outputting distribution of a target appearing in an X-ray image using a learning model, an image quality improvement processing unit, and a display unit. The image quality improvement processing unit is configured to switch, using a learning identification result by a target distribution learning identification unit, between a first image processing mode for performing image quality improvement processing on an X-ray image and a second image processing mode for performing image quality improvement processing on the X-ray image without using the learning identification result.