X-Ray Image Stitching by Border Correlation for Scan Direction Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing x-ray imaging systems face challenges in determining the scan direction and stitching sequence of multiple images accurately, especially when manual movement is involved, leading to difficulties in aligning and combining images due to unknown positional information and potential deviations from a straight line.
Innovation Solution
An apparatus and method using normalized cross-correlation algorithms to compare image regions at borders to determine the scan direction and stitching sequence by calculating similarity values for different positional overlays, allowing for precise alignment and combination of x-ray images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual displacement of the X-ray acquisition system is used, then ease of operation is improved, but measurement precision of detector position deteriorates
Solution Approach 1:
The patent replaces mechanical position tracking (motorized movement with encoder feedback) with an image-based computational method. By using normalized cross-correlation algorithms to compare overlapping regions in sequential X-ray images, the system automatically determines scan direction and stitching sequence without relying on mechanical position data from motorized movement or manual operator input.
2Measurement precision
If motorized movement with position information is used, then measurement precision of detector position is improved, but device complexity increases
Solution Approach 1:
The patent eliminates the need for motorized movement and associated position tracking systems by substituting them with a software-based image correlation approach. The processing unit compares image data from sequential acquisitions using normalized cross-correlation to automatically determine the stitching sequence, thereby reducing mechanical and control system complexity.
Solution Approach 2:
The system performs self-positioning through image analysis. Instead of relying on external motor control systems to track and report position, the X-ray images themselves contain the positional information needed for stitching. The algorithm extracts this information by analyzing overlapping anatomical structures in the images, making the system self-sufficient.
3Adaptability or versatility
If the movement deviates from a straight line, then adaptability to manual operation is improved, but manufacturing precision of image alignment deteriorates
Solution Approach 1:
The patent employs a dynamic image correlation algorithm that can handle non-linear movement paths. The normalized cross-correlation method automatically adapts to the actual movement trajectory by finding the optimal alignment between overlapping image regions, regardless of whether the movement was straight, curved, or irregular. This dynamic approach maintains alignment precision while accommodating manual operation variability.
Data Source
AI summary
The present invention relates to an apparatus (10) for scan direction detection and stitching sequence determination of a plurality of X-ray images, comprising: an input unit (20); a processing unit (30); and an output unit (40). The input unit is configured to provide the processing unit with a first X-ray image acquired by an X-ray image acquisition system, and the first image comprises image data of a patient. The input unit is configured to provide the processing unit with a second X-ray image acquired by the X-ray image acquisition system after it has moved with respect to the patient, and the second image comprises image data of the patient. The processing unit is configured to determine a top similarity value comprising a comparison of at least one region of image data of the patient at and/or adjacent to a top border in the first image with at least one equivalent sized region of image data of the patient at and/or adjacent to a bottom border in the second image. The processing unit is configured to determine a right similarity value comprising a comparison of at least one region of image data of the patient at and/or adjacent to a right border in the first image with at least one equivalent sized region of image data of the patient at and/or adjacent to a left border in the second image. The processing unit is configured to determine a bottom similarity value comprising a comparison of at least one region of image data of the patient at and/or adjacent to a bottom border in the first image with at least one equivalent sized region of image data of the patient at and/or adjacent to a top border in the second image. The processing unit is configured to determine a left similarity value comprising a comparison of at least one region of image data of the patient at and/or adjacent to a left border in the first image with at least one equivalent sized region of image data of the patient at and/or adjacent to a right border in the second image. The processing unit is configured to determine a scan direction and translation distance of the X-ray image acquisition system associated with the movement of the X-ray acquisition system comprising utilization of a maximum of the top, right, bottom or left similarity values and/or determine a combined image formed from the first image and the second image comprising utilization of the maximum of the top, right, bottom or left similarity values. The output unit is configured to output the scan direction and translation distance and/or the combined image.