東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Dense matching and image segmentatio...

O'Connell, David John.

FindBook

Google Book

Amazon

博客來

Dense matching and image segmentation using projective geometry.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Dense matching and image segmentation using projective geometry./
作者:	O'Connell, David John.
面頁冊數:	70 p.
附註:	Source: Masters Abstracts International, Volume: 41-04, page: 1116.
Contained By:	Masters Abstracts International41-04.
標題:	Computer Science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=MQ75800
ISBN:	0612758001

Dense matching and image segmentation using projective geometry.
O'Connell, David John.

Dense matching and image segmentation using projective geometry. - 70 p.

Source: Masters Abstracts International, Volume: 41-04, page: 1116.

Thesis (M.Sc.)--University of Windsor (Canada), 2002.

Dense matching and image segmentation are fundamental image analysis operations. These operations are required by many computer vision applications. Artificial view-synthesis, 3D scene reconstruction, token tracking and augmented reality are examples of applications that rely heavily on these primitives. The speed and accuracy of such applications rely on the quality of the matching and segmentation. As a result, solutions to these problems have been widely researched. However, due to the difficulty of these problems, there is no universal solution.

ISBN: 0612758001Subjects--Topical Terms:

626642
Computer Science.

Dense matching and image segmentation using projective geometry.
LDR:02327nmm 2200289 4500 001 1866212
005 20041220114134.5
008 130614s2002 eng d
020 $a 0612758001
035 $a (UnM)AAIMQ75800
035 $a AAIMQ75800
040 $a UnM $c UnM
100 1 $a O'Connell, David John. $3 1282507
245 1 0 $a Dense matching and image segmentation using projective geometry.
300 $a 70 p.
500 $a Source: Masters Abstracts International, Volume: 41-04, page: 1116.
500 $a Adviser: Bubaker Boufama.
502 $a Thesis (M.Sc.)--University of Windsor (Canada), 2002.
520 $a Dense matching and image segmentation are fundamental image analysis operations. These operations are required by many computer vision applications. Artificial view-synthesis, 3D scene reconstruction, token tracking and augmented reality are examples of applications that rely heavily on these primitives. The speed and accuracy of such applications rely on the quality of the matching and segmentation. As a result, solutions to these problems have been widely researched. However, due to the difficulty of these problems, there is no universal solution.
520 $a Most solutions to these two problems make certain assumptions. First, dense matching and image segmentation are often viewed as separate problems. Second, most image segmentation techniques operate on only a single image. This introduces a technique that simultaneously performs image segmentation and dense matching of planar surfaces in a stereo pair of images. Using three matched points from an arbitrary plane, and four other matched points, a projective mapping, known as a homography, is calculated. This homography is used to iteratively grow a region in both images. The result is a matched and segmented plane.
520 $a Practical tests comparing the computation time of this method to traditional matching techniques are presented. These results are used to motivate the use of the planar technique as a primary step for reducing the overall computation time for dense matching and image segmentation.
590 $a School code: 0115.
650 4 $a Computer Science. $3 626642
690 $a 0984
710 2 0 $a University of Windsor (Canada). $3 1018526
773 0 $t Masters Abstracts International $g 41-04.
790 1 0 $a Boufama, Bubaker, $e advisor
790 $a 0115
791 $a M.Sc.
792 $a 2002
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=MQ75800