View : 649 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor윤정호-
dc.contributor.author정지윤-
dc.creator정지윤-
dc.date.accessioned2017-08-27T12:08:43Z-
dc.date.available2017-08-27T12:08:43Z-
dc.date.issued2017-
dc.identifier.otherOAK-000000137612-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000137612en_US
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/236349-
dc.description.abstractWe present a method for edge detection based on the moving least squares (MLS) algorithm. Since the MLS method is robust to non-smooth function, it is useful for approxion to function containing noisy data. So, the proposed method uses the MLS method to approximate derivative of function. Also, we use the minmod function for edge detection method. When we use high order edge detection methods, it converges to zero for any point away from discontinuities. On the other hand, it produces the more oscillations near the discontinuities. The minmod function applied to edge detection method with various orders can determine true discontinuities from the oscillations by reducing neighborhood oscillations. Our edge detection method can obtain the magnitude of the first derivative continuities as well as the magnitude of jump discontinuities. Also, we present the edge detection algorithm for two dimension case. Some numerical results are provided to demonstrate the ability of the new edge detection method. Including the synthetic image with known ground truth edge, we use various test images. Especially, for a synthetic image with known ground truth edge, the quantative measure, Pratt’s figure of merit, is used to compare the performance of edge detectors.;이 논문은 MLS 근사법을 기반으로 한 edge detection 방법을 소개한다. 이 논문에서 제안한 edge detection 방법은 점프 불연속점과 한번 미분하여 생긴 불연속점들에 대해 적용되어 질 수 있다. 불연속점들을 찾는 과정에서 차수가 증가할 수록 연속인 점들에서 0으로 수렴하는 속도가 증가하는 반면에 불연속점 주변에서 진동들이 발생하게 된다. 이를 보안하기 위해, 우리가 제안하는 방법은 minmod 함수를 적용하여 불연속점 주변에서의 진동들을 감소시켜 진동들로 인해 불연속점을 잘 못 찾는 것을 막아준다. 최종적으로 이를 이용한 알고리즘으로 일차원 함수들을 가지고 테스트하여 잘 작동되는지에 대해 보였으며, 이차원 이미지들에 대해서는 시각적인 결과 뿐만 아니라 수치적인 결과를 제시하여 제안한 방법의 이점에 대해 보였다.-
dc.description.tableofcontents1 Introduction 1 2 Preliminaries 3 2.1 Moving Least Squares approximation 3 2.2 Edge detection method in one dimension 5 2.3 Minmod edge detection method 7 3 Construction of a New Edge Detection Method based on Least-squares constraints 8 3.1 Approximation of derivatives by MLS 9 3.2 Edge detection method based on least-squares constraints 10 4 Experimental Results 14 4.1 Examples in one dimension 15 4.2 Examples in two dimension 20 5 Conclusion 25 References 26 국문초록 28-
dc.formatapplication/pdf-
dc.format.extent1540099 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleConstruction of a New Edge Detection Method based on Least-squares constraints-
dc.typeMaster's Thesis-
dc.format.pageii, 28 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 수학과-
dc.date.awarded2017. 2-
Appears in Collections:
일반대학원 > 수학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE