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

Foreign Element Detection in Chest X...

Zohora, Fatema Tuz.

FindBook

Google Book

Amazon

博客來

Foreign Element Detection in Chest X-ray Images.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Foreign Element Detection in Chest X-ray Images./
作者:	Zohora, Fatema Tuz.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2016,
面頁冊數:	118 p.
附註:	Source: Masters Abstracts International, Volume: 56-04.
Contained By:	Masters Abstracts International56-04(E).
標題:	Computer science. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10278832
ISBN:	9781369762938

Foreign Element Detection in Chest X-ray Images.
Zohora, Fatema Tuz.

Foreign Element Detection in Chest X-ray Images. - Ann Arbor : ProQuest Dissertations & Theses, 2016 - 118 p.

Source: Masters Abstracts International, Volume: 56-04.

Thesis (M.S.)--University of South Dakota, 2016.

In an automated chest X-ray (CXR) screening (to detect pulmonary abnormalities -- Tuberculosis (TB), for instance), the presence of the foreign element (e.g., buttons, coins, medical devices) hinders its performance. In this thesis work, I present novel techniques for detecting circle-like element and medical tubes and devices. At first, I focus on detecting circle-like foreign element (both with/without considering the lung region) in the chest X-ray images. I start with pre-processing steps to enhance the CXR images and then detect foreign element. My proposed methods can be categorized into two types: 1) circular assumption-based and 2) training-based. In the first method, candidate selection followed by circular Hough transform (CHT), compute edge images using several different edge detection algorithms and then apply morphological operations to select candidate regions (image segmentation) in the lung region. Finally, CHT is used to detect the circular foreign element. In the second method, I perform the normalized cross-correlation (NCC) followed by an unsupervised clustering. In all tests, both techniques show good performance for a large number of CXR images. I also compare the performance of the proposed techniques with existing methods in the literature (Viola-Jones and CHT). Our methods excelled in performance both in terms of detection accuracy (precision, recall, and F1 score) and computational time. At the output, precision, recall, and F1 score are 90%(96%), 93%(90%), and 91%(92%) from circular assumption-based technique (training- based method). Further, I focus on medical device (with/without tubes) detection in CXRs using training-based, and the results are encouraging.

ISBN: 9781369762938Subjects--Topical Terms:

523869
Computer science.

Foreign Element Detection in Chest X-ray Images.
LDR:02569nmm a2200289 4500 001 2126651
005 20171128150729.5
008 180830s2016 ||||||||||||||||| ||eng d
020 $a 9781369762938
035 $a (MiAaPQ)AAI10278832
035 $a AAI10278832
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zohora, Fatema Tuz. $3 3288757
245 1 0 $a Foreign Element Detection in Chest X-ray Images.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2016
300 $a 118 p.
500 $a Source: Masters Abstracts International, Volume: 56-04.
500 $a Adviser: Santosh KC.
502 $a Thesis (M.S.)--University of South Dakota, 2016.
520 $a In an automated chest X-ray (CXR) screening (to detect pulmonary abnormalities -- Tuberculosis (TB), for instance), the presence of the foreign element (e.g., buttons, coins, medical devices) hinders its performance. In this thesis work, I present novel techniques for detecting circle-like element and medical tubes and devices. At first, I focus on detecting circle-like foreign element (both with/without considering the lung region) in the chest X-ray images. I start with pre-processing steps to enhance the CXR images and then detect foreign element. My proposed methods can be categorized into two types: 1) circular assumption-based and 2) training-based. In the first method, candidate selection followed by circular Hough transform (CHT), compute edge images using several different edge detection algorithms and then apply morphological operations to select candidate regions (image segmentation) in the lung region. Finally, CHT is used to detect the circular foreign element. In the second method, I perform the normalized cross-correlation (NCC) followed by an unsupervised clustering. In all tests, both techniques show good performance for a large number of CXR images. I also compare the performance of the proposed techniques with existing methods in the literature (Viola-Jones and CHT). Our methods excelled in performance both in terms of detection accuracy (precision, recall, and F1 score) and computational time. At the output, precision, recall, and F1 score are 90%(96%), 93%(90%), and 91%(92%) from circular assumption-based technique (training- based method). Further, I focus on medical device (with/without tubes) detection in CXRs using training-based, and the results are encouraging.
590 $a School code: 0203.
650 4 $a Computer science. $3 523869
650 4 $a Medical imaging. $3 3172799
690 $a 0984
690 $a 0574
710 2 $a University of South Dakota. $b Computer Science. $3 3288758
773 0 $t Masters Abstracts International $g 56-04(E).
790 $a 0203
791 $a M.S.
792 $a 2016
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10278832