Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information
碩士 === 國立中正大學 === 資訊工程所 === 93 === In conventional quantitative analysis of dual phase Tl-201 SPECT dataset, the physician objectively and manually marks the regions of the lesion and the background on a single 2-D image to evaluate their retention index. The selections of the tumor image and the co...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2005
|
Online Access: | http://ndltd.ncl.edu.tw/handle/18585175302181031207 |
id |
ndltd-TW-093CCU05392017 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-093CCU053920172015-10-13T10:45:04Z http://ndltd.ncl.edu.tw/handle/18585175302181031207 Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information 結合電腦斷層掃描之兩階段Tl-201單光子斷層掃描之肺部腫瘤電腦輔助診斷 Su-yun Wang 王舒韻 碩士 國立中正大學 資訊工程所 93 In conventional quantitative analysis of dual phase Tl-201 SPECT dataset, the physician objectively and manually marks the regions of the lesion and the background on a single 2-D image to evaluate their retention index. The selections of the tumor image and the corresponding background affect the reliability and usability of retention index (RI) used for classifying the lesions. Moreover, the CT information is used for assisting in selecting an appropriate region of the background. In order to solve the disadvantages of retention index method, we use 3-D objects to calculate the retention index and investigate a computer-aided diagnosis system which could automatically find the lung lesions from the dual phase Tl-201 SPECT datasets and distinguish the benign lesions from the malignant ones. In this paper, we replace the background in the conventional retention index formula with the heart, and the new retention index with heart RIHA is defined and used for quantitative analysis. The proposed CAD system uses the optimal threshold which proposed by Otsu’s to obtain the suspected regions of the heart and the lesions, and then the heart is decided by several conditions. According to the heart, the dual phase Tl-201 SPECT datasets and CT data set are synchronized, and then the 3-D objects with the related 2-D ROIs are built. Finally the CAD system evaluates the RIHA of suspected 3-D objects and distinguishes the benign lesions from the malignant ones according to the RIHA. With the experiments of 13 benign and 15 malignant cases, the accuracy of this method reaches 85.71%. Ruey-feng Chang 張瑞峰 2005 學位論文 ; thesis 61 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立中正大學 === 資訊工程所 === 93 === In conventional quantitative analysis of dual phase Tl-201 SPECT dataset, the physician objectively and manually marks the regions of the lesion and the background on a single 2-D image to evaluate their retention index. The selections of the tumor image and the corresponding background affect the reliability and usability of retention index (RI) used for classifying the lesions. Moreover, the CT information is used for assisting in selecting an appropriate region of the background. In order to solve the disadvantages of retention index method, we use 3-D objects to calculate the retention index and investigate a computer-aided diagnosis system which could automatically find the lung lesions from the dual phase Tl-201 SPECT datasets and distinguish the benign lesions from the malignant ones. In this paper, we replace the background in the conventional retention index formula with the heart, and the new retention index with heart RIHA is defined and used for quantitative analysis. The proposed CAD system uses the optimal threshold which proposed by Otsu’s to obtain the suspected regions of the heart and the lesions, and then the heart is decided by several conditions. According to the heart, the dual phase Tl-201 SPECT datasets and CT data set are synchronized, and then the 3-D objects with the related 2-D ROIs are built. Finally the CAD system evaluates the RIHA of suspected 3-D objects and distinguishes the benign lesions from the malignant ones according to the RIHA. With the experiments of 13 benign and 15 malignant cases, the accuracy of this method reaches 85.71%.
|
author2 |
Ruey-feng Chang |
author_facet |
Ruey-feng Chang Su-yun Wang 王舒韻 |
author |
Su-yun Wang 王舒韻 |
spellingShingle |
Su-yun Wang 王舒韻 Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
author_sort |
Su-yun Wang |
title |
Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
title_short |
Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
title_full |
Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
title_fullStr |
Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
title_full_unstemmed |
Computer-aided Diagnosis of Lung Lesions for Dual Phase Tl-201 SPECT Images using CT Information |
title_sort |
computer-aided diagnosis of lung lesions for dual phase tl-201 spect images using ct information |
publishDate |
2005 |
url |
http://ndltd.ncl.edu.tw/handle/18585175302181031207 |
work_keys_str_mv |
AT suyunwang computeraideddiagnosisoflunglesionsfordualphasetl201spectimagesusingctinformation AT wángshūyùn computeraideddiagnosisoflunglesionsfordualphasetl201spectimagesusingctinformation AT suyunwang jiéhédiànnǎoduàncéngsǎomiáozhīliǎngjiēduàntl201dānguāngziduàncéngsǎomiáozhīfèibùzhǒngliúdiànnǎofǔzhùzhěnduàn AT wángshūyùn jiéhédiànnǎoduàncéngsǎomiáozhīliǎngjiēduàntl201dānguāngziduàncéngsǎomiáozhīfèibùzhǒngliúdiànnǎofǔzhùzhěnduàn |
_version_ |
1716832016861757440 |