SIFT and Otsu matching based colored eyeground image splicing method

Abstract

The invention relates to a SIFT and Otsu matching based colored eyeground image splicing method. The method is realized by that (1) a green channel of a reference eyeground image and a target eyeground image is extracted; (2) the green channel is preprocessed via contract limited adaptive histogram equalization and two-dimensional Gaussian filtering to enhance the contract between vessels and tissue background; (3) SIFT characteristics of the image are extracted; (4) characteristic points of the two images are matched initially via a nearest neighbor and a next nearest neighbor algorithm; (5)an Otsu algorithm is used to screen the initial matching points to obtain correct matching characteristic point pairs; (6) a random sampling consistency algorithm is used to calculate an affine transformation matrix to carry out space transformation o the images and complete registration; and (7) the registering images are fused in a maximum value algorithm, and splicing is completed. According tothe invention, SIFT is combined with the Otsu method, the disadvantage of non-ideal registering effect caused by mismatching points is overcome, and the method has significance in the field of diagnosing diabetes and pathology degree thereof.

Description

technical field [0001] The invention relates to a mosaic method, which can mosaic fundus images of different lesions, belongs to the technical field of image processing, and can be applied to mosaic of multiple fundus images and registration of multi-mode fundus images. Background technique [0002] Fundus images are an important basis for ophthalmology clinical diagnosis. In the screening of diabetic retinopathy, due to the limited field of view of the camera when obtaining images, a single image can only reflect the local information of the fundus, which is difficult to meet the needs of doctors. In order to obtain fundus images with a large field of view, it is necessary to image the same fundus from different angles, and clinical doctors generally refer to multiple images taken at the same time for diagnosis. If multiple images can be fused into a complete fundus image, it is of great significance for doctors to make judgments. [0003] At present, fundus image stitchi...

AI Technical Summary

Problems solved by technology

But there are still some problems. The region-based method is sensitive to the gray scale of the image. Since there are a large number of uniform and similar regions in the fundus image, the avascular region occupies most of the fundus area. The region-based method is difficult to meet the requirements in the fundus image mosaic; The feature method is based on the matching features between images, and extracts the points, lines, contours and other features of the image as matching features, and uses the similarity measurement value between the features to match and complete the splicing. Most of the current good splicing results use The crossing branch points of blood vessels are used as matching features, and the mode of blood vessel tracking and feature extraction is used to complete image stitching by using the maximum overlap of blood vessel information. The above methods all rely on the extraction of blood vessel structures and intersection points of blood vessels. When the fundus lesion is serious and the blood vessel structure is blurred, The performance of the method will be greatly reduced. Feature point matching is a key step in image registration. Feature description uses the width and direction of blood vessels around feature points, but these features are not robust in matching, and the feature description of feature points is not always differentiated from each other

Images