109 0

Displaying low light high dynamic range images using fuzzy set theory

Title
Displaying low light high dynamic range images using fuzzy set theory
Author
박미선
Alternative Author(s)
Piao,Mei Xian
Advisor(s)
정제창
Issue Date
2017-08
Publisher
한양대학교
Degree
Master
Abstract
High dynamic range(HDR) imaging has been developed for almost 30 years. This technique will be used in diverse areas such as medical imaging and UHD TV. Because of different bit depth, HDR images needs to be displayed by high dynamic range monitors. However, most of users still rely on low dynamic range(LDR) monitors. One of the available approaches is tone mapping operators which map high dynamic range contents to LDR monitors. Until now plenty of tone mapping operators have been presented. Some of them have a good performance to deal with most of HDR contents but not low light contents. In this thesis, we present a tone mapping technique for low light HDR scene, which is using edge preserving filtering framework and fuzzy enhancement. The HDR image is decomposed into base layer and detail layer. The base layer reduces contrast which is obtained by edge preserving filter such as guided filter. After filtering, we will get edge preserved base layer which usually has large scale variation. Detail layer is the residual of the original signal subtracting base layer. We improve performance of the low light image using fuzzy enhancement. In general, low light HDR images have unusual histograms. The traditional tone mapped images neither lose detail nor sacrifice the balance of the brightness. We find that using fuzzy set theory can fix this kind of problem. This thesis contributes a perceptual tone mapping operator which produces a better visual appearance than traditional tone mapping operators. This algorithm is especially work well on the low light contents. More specifically, the algorithm increases darker region brightness and decreases brighter region brightness. The final result does change too much original histogram, at the same time, almost keeping the structure of the original histogram.
URI
http://dcollection.hanyang.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000102162http://hdl.handle.net/20.500.11754/33603
Appears in Collections:
GRADUATE SCHOOL[S](대학원) > ELECTRONICS AND COMPUTER ENGINEERING(전자컴퓨터통신공학과) > Theses (Ph.D.)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE