Eyeball Image Registration and Fusion Based on SIFT+RANSAC Algorithm

Abstract

At present, there are still many problems: the related technologies for registration and fusion of medical images are not mature enough, and the time for feature extraction and matching is too long; the matching points are prone to redundancy, and the image fusion has gaps. Or it may lead to blurring and other phenomena. During the medical image analysis, it is necessary to put together several images of the same patient for analysis, thereby obtaining comprehensive information of the patient in various aspects and improving the level of medical diagnosis and treatment. To quantitatively analyze several different images, we must first solve the strict alignment problem of these images. This is what we call the registration of images. Image fusion technology combines various images to display their own information on the same image, providing multi-data and multi-information images for clinical medical diagnosis. This becomes an application-critical technology, and accurate and efficient image matching criterion is a key and difficult point. Therefore, image registration and fusion technology are of great significance both in computer vision and in clinical medical diagnosis. Through the retrieval of existing related technologies, this paper used a medical image registration and fusion method based on scale-invariant feature transform (SIFT) and random sample consensus (RANSAC) algorithm, which is used to solve the following technical problems of medical image registration and fusion: feature points’ extraction and matching time is too long; the matching correction problem; and the fusion of the image is prone to gaps or blurs. The method of this paper is suitable for clinically detecting the disease information of the same patient, and more valuable information can be obtained through registration and fusion.