Radiographic Detector Edge Boundary Definition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital radiography systems face complexity and cost issues in defining a region of interest (ROI) for asymmetrical imaging due to the need for multiple position-sensing components, which can lead to calibration challenges and errors.
Innovation Solution
A method that uses image processing to detect collimator blade edges by accumulating pixel values along specific vectors and identifying threshold values to define image boundaries, eliminating the need for costly and error-prone orientation sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple position-sensing components are used to detect collimator orientation angles, then the region of interest can be defined for asymmetrical imaging, but the device complexity and cost increase
Solution Approach 1:
The patent replaces mechanical position-sensing components with a computational image processing system. The method uses digital image data from the detector and applies projection profile analysis to automatically detect collimator blade edges and define the ROI, eliminating the need for tilt sensors and position reporting components while maintaining the capability to handle asymmetrical imaging configurations
Solution Approach 2:
The imaging system uses its own captured image data to automatically define the region of interest. By analyzing the projection profiles of the captured image, the system self-determines the collimator blade edges and ROI boundaries without requiring external sensing components, making the system self-sufficient in defining the imaged area
2Measurement precision
If multiple position-sensing components are used to detect collimator orientation, then ROI definition is achieved, but calibration challenges and errors increase
Solution Approach 1:
The patent eliminates mechanical sensing components that require calibration by using a purely computational approach. The method processes digital image data through projection profile analysis to detect collimator blade edges, removing the calibration chain entirely and replacing it with algorithm-based detection that is inherently more reliable and free from calibration drift
Solution Approach 2:
The system uses feedback from the captured image data itself to define the ROI. By analyzing the actual projection profiles obtained from the imaged area, the system automatically adjusts and defines the ROI boundaries based on real image content, creating a closed-loop system that is self-correcting and does not rely on pre-calibration data
3Loss of information
If the entire image area is processed, then complete image data is available, but image processing time increases
Solution Approach 1:
The patent extracts only the diagnostically relevant portion of the image by automatically defining and processing only the ROI. The method identifies collimator blade edges in the full image and then extracts and processes only the region within those boundaries, separating the useful image data from the unnecessary background areas to reduce processing time while maintaining diagnostic information completeness
Solution Approach 2:
The patent segments the image into the ROI and non-ROI portions based on detected collimator blade edges. By dividing the image data into relevant and irrelevant sections and processing only the relevant segment, the system reduces the total processing load and time while preserving all diagnostically important information within the defined ROI
4Area of stationary object
If the entire image area is displayed, then complete image is visible, but background glare interferes with image quality
Solution Approach 1:
The patent extracts and isolates the ROI from the full image area, displaying only the relevant portion. By removing the background areas outside the collimator-defined boundaries, the system eliminates the source of background glare and interference while maintaining the visibility of the complete diagnostically relevant image content
Solution Approach 2:
The patent applies different treatment to different regions of the image by defining a specific ROI with distinct boundaries. The region inside the collimator blade edges receives full processing and display, while the exterior regions are excluded or suppressed, applying local quality enhancement to the relevant area and eliminating glare from irrelevant background areas
Data Source
AI summary
A method for defining edge boundaries of an image formed on a flat-panel radiographic detector. The method includes obtaining digital image pixel values from the radiographic detector; obtaining dimensional information about a collimator in the radiographic signal path; analyzing the image to detect one or more collimator blade edges; and processing the digital image pixel values at least once to detect remaining collimator blade edges. The method accumulates image pixel values along each of a plurality of lines in a first direction parallel to a first vector that extends across the image, to obtain a first profile value for each line in said first direction, thereby forming an ordered set of first profile values. The method then identifies first and second threshold values in the set of first profile values, wherein the paired first and second threshold values indicate an edge boundary corresponding to a collimator blade projection along the first direction.