The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance

碩士 === 元智大學 === 電機工程學系 === 106 === Image matting purposes to cut the foreground region from an image by calculating pixel’s color value and estimating alpha value. In the past ten years, lots of papers on image matting have been published. Most of them obtain lower difference by using laplacian matt...

Full description

Bibliographic Details
Main Authors: Tsung-Wei Wu, 吳崇維
Other Authors: Yung-Sheng Chen
Format: Others
Language:zh-TW
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/v5nst8
id ndltd-TW-106YZU05442028
record_format oai_dc
spelling ndltd-TW-106YZU054420282019-10-31T05:22:13Z http://ndltd.ncl.edu.tw/handle/v5nst8 The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance 基於EMD改善共享採樣切割演算法之研究 Tsung-Wei Wu 吳崇維 碩士 元智大學 電機工程學系 106 Image matting purposes to cut the foreground region from an image by calculating pixel’s color value and estimating alpha value. In the past ten years, lots of papers on image matting have been published. Most of them obtain lower difference by using laplacian matting model. However, those algorithms’ laplacian matrix has too many elements, it takes a lot of time to solve the linear equations. In this paper, we do not want to spend too much time in the computing process, so choosing Shared Matting as our basic algorithm because it’s low cost time. And using Earth Mover’s Distance and simple laplacian matting to find each pixels’ alpha value. The results obtained by our method have lower difference than Shared Matting, and the computing time is within our expectation. Yung-Sheng Chen 陳永盛 2018 學位論文 ; thesis 55 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 碩士 === 元智大學 === 電機工程學系 === 106 === Image matting purposes to cut the foreground region from an image by calculating pixel’s color value and estimating alpha value. In the past ten years, lots of papers on image matting have been published. Most of them obtain lower difference by using laplacian matting model. However, those algorithms’ laplacian matrix has too many elements, it takes a lot of time to solve the linear equations. In this paper, we do not want to spend too much time in the computing process, so choosing Shared Matting as our basic algorithm because it’s low cost time. And using Earth Mover’s Distance and simple laplacian matting to find each pixels’ alpha value. The results obtained by our method have lower difference than Shared Matting, and the computing time is within our expectation.
author2 Yung-Sheng Chen
author_facet Yung-Sheng Chen
Tsung-Wei Wu
吳崇維
author Tsung-Wei Wu
吳崇維
spellingShingle Tsung-Wei Wu
吳崇維
The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
author_sort Tsung-Wei Wu
title The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
title_short The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
title_full The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
title_fullStr The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
title_full_unstemmed The Study of Improving the Shared Sampling Method for Alpha Matting Based on Earth Mover's Distance
title_sort study of improving the shared sampling method for alpha matting based on earth mover's distance
publishDate 2018
url http://ndltd.ncl.edu.tw/handle/v5nst8
work_keys_str_mv AT tsungweiwu thestudyofimprovingthesharedsamplingmethodforalphamattingbasedonearthmoversdistance
AT wúchóngwéi thestudyofimprovingthesharedsamplingmethodforalphamattingbasedonearthmoversdistance
AT tsungweiwu jīyúemdgǎishàngòngxiǎngcǎiyàngqiègēyǎnsuànfǎzhīyánjiū
AT wúchóngwéi jīyúemdgǎishàngòngxiǎngcǎiyàngqiègēyǎnsuànfǎzhīyánjiū
AT tsungweiwu studyofimprovingthesharedsamplingmethodforalphamattingbasedonearthmoversdistance
AT wúchóngwéi studyofimprovingthesharedsamplingmethodforalphamattingbasedonearthmoversdistance
_version_ 1719284685899563008