Progressive Image Registration for Brain Imaging Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional medical image registration methods suffer from low registration accuracy due to the large structural differences between medical images, which are exacerbated by poor image quality and difficulty in feature extraction, leading to unsatisfactory results.
Innovation Solution
A two-stage medical image registration method based on progressive images (PIs) and the Speeded-up Robust Features (SURF) algorithm, where multiple intermediate progressive images are generated by merging reference and floating images, and feature points are matched using affine transformations and bicubic interpolation to achieve accurate registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image registration methods (SURF algorithm) are used to directly register floating image and reference image, then the registration process is simple, but the registration accuracy is low due to large structural differences and poor image quality
Solution Approach 1:
The patent divides the registration process into two distinct stages: coarse registration and fine registration. Coarse registration uses the SURF algorithm to achieve initial alignment, while fine registration applies B-spline transformation for precision refinement. This segmentation allows each stage to focus on specific requirements, improving overall accuracy without overwhelming complexity.
Solution Approach 2:
The patent performs preliminary coarse registration using the SURF algorithm before conducting fine registration. By pre-aligning the images through affine transformation in the coarse stage, the subsequent fine registration stage operates on pre-aligned images, making feature extraction and matching more effective and accurate.
2Productivity
If the SURF algorithm is used for feature-based registration, then the computation speed is improved, but the registration accuracy deteriorates due to difficulty in feature extraction from poor quality images
Solution Approach 1:
The patent segments the registration task into two phases: coarse registration using SURF for speed, and fine registration using B-spline for accuracy. This allows the system to benefit from the computational efficiency of SURF while compensating for its accuracy limitations through the subsequent refinement stage.
Solution Approach 2:
The patent introduces an intermediate affine transformation result as a mediator between the original images and the final registered output. The coarse registration produces an intermediate aligned image that serves as a better basis for fine registration, bridging the gap between speed and accuracy requirements.
3Reliability
If direct registration of floating image and reference image is performed, then the process is efficient, but the registration results are unsatisfactory due to large structural differences
Solution Approach 1:
The patent divides registration into coarse and fine stages, where each stage performs a specific function. Coarse registration handles large structural differences through affine transformation, while fine registration refines the alignment. This segmentation makes the overall process more reliable despite increased time investment.
Solution Approach 2:
The patent performs preliminary coarse alignment to reduce large structural differences before attempting fine registration. By pre-processing the images to achieve rough alignment, the subsequent fine registration can focus on precision adjustments, improving overall result quality.
Data Source
AI summary
A two-stage medical image registration method based on progressive images (PIs) to solve the technical problem of low registration accuracy of traditional image registration methods includes: merging a reference image with a floating image to generate multiple intermediate PIs; registering, by a speeded-up robust features (SURF) algorithm and an affine transformation, the floating image with the intermediate PIs to acquire coarse registration results; registering, by the SURF algorithm and the affine transformation, the reference image with the coarse registration results to acquire fine registration results; and comparing the fine registration results of the intermediate PIs, which are acquired by iteration, and selecting an optimal registration result as a final registration image. The method can achieve multimodal registration for brain imaging with MI, NCC, MSD, and NMI superior to those of the existing registration algorithms. The method effectively improves the registration accuracy through the progressive medical image registration strategy.