Weighted Pixel Registration for Medical Image Stitching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing technologies face challenges in stitching overlapped medical images, such as MRI scans, due to limitations in imaging apparatus size, resulting in incomplete images that require stitching of multiple scans, which can lead to misalignment and poor transition between overlapped areas.
Innovation Solution
An image processing apparatus and method that performs registration on two overlapped images using a first registration device to match pixels in the overlapped area, with an output pixel location determination device calculating the location of output pixels based on weighted locations of matched pixels, and an output pixel value determination device calculating pixel values, ensuring a smooth transition between overlapped and non-overlapped areas by using nonlinear weight functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple scans are performed to image the whole object, then complete coverage is achieved, but misalignment and poor transition occur between overlapped areas
Solution Approach 1:
The patent replaces mechanical alignment methods with image processing-based registration techniques. By using pixel matching and transformation calculations between overlapped images, the system achieves precise alignment without relying on mechanical positioning accuracy, thus resolving the contradiction between coverage area and alignment precision.
Solution Approach 2:
The patent transforms the alignment problem from a spatial positioning problem to a parameter transformation problem. By calculating transformation parameters (translation, rotation, scaling) based on pixel matching in overlapped areas, the system achieves precise registration, resolving the contradiction between complete coverage and alignment precision.
2Manufacturing precision
If registration is performed on the whole image, then comprehensive alignment is achieved, but calculation complexity increases
Solution Approach 1:
The patent divides the image processing task into two segments: registration is performed only on the overlapped area to determine transformation parameters, while the non-overlapped areas are directly copied without registration. This segmentation reduces calculation complexity while maintaining alignment precision in the critical overlapped regions.
Solution Approach 2:
The patent applies different processing quality to different regions: high-quality registration is applied only to the overlapped area where precision is critical, while the non-overlapped area uses simpler direct copying. This local quality approach maintains necessary alignment precision while reducing overall calculation complexity.
3Device complexity
If simple copying is used for overlapped areas, then calculation complexity is reduced, but smooth transition between areas is lost
Solution Approach 1:
The patent uses transformation parameters calculated from the overlapped area as an intermediary to guide the copying process. These parameters act as a mediator that ensures the copied non-overlapped area aligns smoothly with the overlapped area, maintaining transition smoothness while keeping the overall process simple.
Data Source
AI summary
An image processing apparatus may include: a first registration device for performing, by taking a first input image of two overlapped input images having an overlapped area as a reference image, a first registration on a second input image to find, in the second input image, a second pixel which is matched with each first pixel located in the overlapped area of the reference image; an output pixel location determination device for calculating a location of an output pixel which is located in the overlapped area of the output image and corresponds to the first pixel, the locations of the first and second pixels being respectively weighted, and the shorter the distance from the first pixel to a non-overlapped area of the reference image is, the greater a weight of the location of the first pixel is; and an output pixel value determination device for calculating a pixel value.