Font Size: a A A

Edge Structure Preserving Image Denosing Based On Double B-splines

Posted on:2020-10-18Degree:MasterType:Thesis
Country:ChinaCandidate:X L LuFull Text:PDF
GTID:2428330578974178Subject:Statistics
Abstract/Summary:PDF Full Text Request
Image denoising is the basic step in image analysis,such as image restoration,image calibration,image extraction,image enhancement,image modeling,etc.The effect of denoising directly affects the effect of the subsequent steps,so image denois-ing has always been a long-lasting subject in the field of images.And the structures in the image,such as the edges,always have very important physical meanings.For ex-ample,in the analysis of MRI,the edges of the image are often the boundary between gray matter,white matter,and cerebrospinal fluid.Once the boundary processing is not good,it will affect the tissue segmentation in the later stage.The amount of cal-culation of some existing point-by-point denoising algorithms based on local linear kernel estimation will become huge.If taking the image as a surface,image denoising is essentially a surface fitting problem.In this paper,an image preservation structure denoising algorithm based on B-spline is proposed.Firstly,the jump point recognition algorithm on the curve is extended to the two-dimensional surface.The method of identifying the true and false jump points including image opening operation,hypothesis testing and clustering is considered.Then,consider the jump curve interpolation,image segmentation prob?lem,and finally the double B-spline are used to fit each smooth part after segmenta-tion,which makes good use of the advantages of B-spline's speed and performance on fitting smooth surfaces.This paper also compares the proposed algorithm with many existing algorithms on several different image models,which proves the effectiveness of our proposed method.
Keywords/Search Tags:Image denoising, Jump curve, Image segmentation, Double B-spline, Surface fitting
PDF Full Text Request
Related items