A Study on the Description of Texture Images

碩士 === 國立臺灣科技大學 === 電機工程系 === 92 === Abstract Textured image research is a topic worth many effort. In this issue, there is generally no widely accepted way to describe a textured image at present. The usual way has to depend on the property of texture and the domain of the application....

Full description

Bibliographic Details
Main Authors: Li,Yen Hung, 李彥宏
Other Authors: Sheu,Hsin Teng
Format: Others
Language:zh-TW
Published: 2004
Online Access:http://ndltd.ncl.edu.tw/handle/37560617126415431483
id ndltd-TW-092NTUST442063
record_format oai_dc
spelling ndltd-TW-092NTUST4420632015-10-13T13:28:04Z http://ndltd.ncl.edu.tw/handle/37560617126415431483 A Study on the Description of Texture Images 紋理影像描述與探討 Li,Yen Hung 李彥宏 碩士 國立臺灣科技大學 電機工程系 92 Abstract Textured image research is a topic worth many effort. In this issue, there is generally no widely accepted way to describe a textured image at present. The usual way has to depend on the property of texture and the domain of the application. In this research, we try to analyze a type of artificially generated texture images. Based on the feature of texture element combination we use hierarchical modeling and further improve the hierarchical modeling using the projective hierarchical modeling. This model could describe the texture more efficiency. The regular texture image is very suitable to analyze in frequency way since texture element is in fixed placement and shows strong periodicity. We use the Fourier transform to analyze the frequency content of a texture image. We can get the frequency distribution in each direction and get the period of the texture. We can also get the displacement of the texture image by observe the variation of phase. At the end of the research we show that the period calculated by observing the spectrum is very close to the period of that is actually measured. Sheu,Hsin Teng 許新添 2004 學位論文 ; thesis 0 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 碩士 === 國立臺灣科技大學 === 電機工程系 === 92 === Abstract Textured image research is a topic worth many effort. In this issue, there is generally no widely accepted way to describe a textured image at present. The usual way has to depend on the property of texture and the domain of the application. In this research, we try to analyze a type of artificially generated texture images. Based on the feature of texture element combination we use hierarchical modeling and further improve the hierarchical modeling using the projective hierarchical modeling. This model could describe the texture more efficiency. The regular texture image is very suitable to analyze in frequency way since texture element is in fixed placement and shows strong periodicity. We use the Fourier transform to analyze the frequency content of a texture image. We can get the frequency distribution in each direction and get the period of the texture. We can also get the displacement of the texture image by observe the variation of phase. At the end of the research we show that the period calculated by observing the spectrum is very close to the period of that is actually measured.
author2 Sheu,Hsin Teng
author_facet Sheu,Hsin Teng
Li,Yen Hung
李彥宏
author Li,Yen Hung
李彥宏
spellingShingle Li,Yen Hung
李彥宏
A Study on the Description of Texture Images
author_sort Li,Yen Hung
title A Study on the Description of Texture Images
title_short A Study on the Description of Texture Images
title_full A Study on the Description of Texture Images
title_fullStr A Study on the Description of Texture Images
title_full_unstemmed A Study on the Description of Texture Images
title_sort study on the description of texture images
publishDate 2004
url http://ndltd.ncl.edu.tw/handle/37560617126415431483
work_keys_str_mv AT liyenhung astudyonthedescriptionoftextureimages
AT lǐyànhóng astudyonthedescriptionoftextureimages
AT liyenhung wénlǐyǐngxiàngmiáoshùyǔtàntǎo
AT lǐyànhóng wénlǐyǐngxiàngmiáoshùyǔtàntǎo
AT liyenhung studyonthedescriptionoftextureimages
AT lǐyànhóng studyonthedescriptionoftextureimages
_version_ 1717736590614200320