Abstract:Medical image registration is a preprocessing stage in a variety of applications such as change detection and mosaicking. Matching is one of the essential steps in image registration. The Scale invariant feature transform (SIFT) is an applicable algorithm used in image matching problems. In all matching methods, including the SIFT algorithm, incorrect matches are created due to the soft-tissue and similar patterns in medical images. Various methods have been suggested to eliminate incorrect matches, each of which has its drawbacks. For example, the Mean-based adaptive RANSAC (MAR) uses the adaptive transformation model to eliminate incorrect matches. This method effectively eliminates incorrect matches against geometric deviations, but it cannot remove incorrect matches of soft-tissue images. The Graph Transformation Matching (GTM) method, on the other hand, is another important and common method in eliminating incorrect matches. This uses the spatial relationships of the image to eliminate incorrect matches, but is not suitable for removing incorrect matches in images with geometric and noise distortions. However, the literature suggests no method that alleviates these problems simultaneously and that works well in all image types. In this paper, a new method based on the hybrid GTM-Mean-based adaptive RANSAC is suggested. In the proposed method, MAR and GTM algorithms are combined innovatively to maintain the advantages of both methods and to eliminate their weaknesses. In this strategy, if two methods have the same result, it is considered as the final decision. In the opposite case, the final decision is made based on the root mean square error between candidate matching points. Simulation results confirm the suggested method's superiority over standard medical databases such as those of retina and brain compared to classic methods in terms of RMSE and precision. The proposed method has increased matching precision by 37% in average, and reduced RMSE by 40% in average.

Minimum Spanning Tree Integrated with Gradient Information for Medical Image Registration

IMAGE REGISTRATION BASED ON MAXIMIZATION OF GRADIENT CODE MUTUAL INFORMATION

Minimum Spanning Tree Fusing Uniform Sub-Sampling Points and High-Dimensional Features for Medical Image Registration

Minimum Spanning Tree Hierarchically Fusing Multi-feature Points and High-Dimensional Features for Medical Image Registration.

Medical Image Elastic Registration Based on Contour and Feature Points

Entropic graph estimation integrated with SIFT features for medical image non-rigid registration

Minimum Spanning Tree Fusing Multi-Salient Points Hierarchically For Multi-Modality Image Registration

Research of Non-rigid Medical Image Registration Based on Gradient Information

DoG keypoints medical image registration based on minimum spanning tree

Local Registration of Medical Images Using Feature Point and Intensity Information

Complementary scale space keypoint medical image registration based on Rényi entropy

Multi-Modality Image Registration with Gradient Orientation Information Based on Entropic Spanning Graph

Improved watershed transform for medical image segmentation using prior information

An Medical Image Registration Approach Using Improved Hausdorff Distance Combined with Particle Swarm Optimization

Multi-Modality Medical Image Registration Based on Regional Joint Rényi Entropy

A hybrid registration of medical images using intensity information and landmark points

A Local Registration Approach Of Medical Images With Niche Genetic Algorithm

A new hybrid strategy in medical image registration based on graph transformation matching and mean-based RANSAC algorithms

Color Fundus Photograph Registration Based on Feature and Intensity for Longitudinal Evaluation of Diabetic Retinopathy Progression

Multi-Step 2D/3D Image Registration in Image-Guided Spine Surgery

Edge Preserving Multi-Modal Registration Based On Gradient Intensity Self-Similarity