Repository logo
  • English
  • 中文
  • Log In
    New user? Click here to register.Have you forgotten your password?
Repository logo
    Communities & Collections
    Research Outputs
    Fundings & Projects
    People
    Organizations
    Statistics
  • English
  • 中文
  • Log In
    New user? Click here to register.Have you forgotten your password?
  1. Home
  2. College of Health Care and Management/健康管理學院
  3. Department of Medical Informatics/醫學資訊學系
  4. 智慧型乳房腫瘤偵測系統使用迭代模糊線性判別分析
 
  • Details
Options

智慧型乳房腫瘤偵測系統使用迭代模糊線性判別分析

Date Issued
2012
Author(s)
卓育順
Yu-Shun
Advisor
秦群立
URI
https://handle.ncl.edu.tw/11296/68zt4x
https://scholars.csmu.edu.tw/handle/123456789/21161
Abstract
乳癌可以說是女性癌症中最常見的一種,根據世界衛生組織指出,每年全球會有乳癌好發年齡在40-55歲之間的120萬個女性被診斷出有罹患乳癌,近年來越來越被重視在診斷及更有效的診斷方式來提高診斷正確率。由於乳癌腫瘤影像,在灰階值亮度會比其它正常組織為高,且紋理組織非常明顯的兩個特徵,所以本論文針對這兩個特徵提出一套的演算法。在前處理的部分我們首先利用影像投影法對乳房的Mammography影像去除其中不必要的影像區域,接著將影像分割成許多大小相同的區塊,每個區塊的大小皆符合特徵抽取時所需要的區塊大小,然後針對區塊做偵測,若區塊內的像素值全部為零,該區塊我們就不做特徵擷取,這就表示區塊中並未出現乳房的資訊,剩下的區塊就變成了ROI區塊,這些ROI區塊將會做為特徵抽取的區塊,接著是特徵值擷取的部分我們採用具有代表紋理能量特徵的Laws’ Mask特徵、代表灰階值動量的Histogram、共生矩陣及代表灰階分布的Modification Average Distance特徵做為用來偵測乳房腫瘤的四大特徵,我們針對每一個ROI影像區塊去抽取出這四大類特徵,但並非四大類特徵中的所有特徵我們都會使用,依據先期實驗發現,使用Laws’ Mask特徵和Average Distance特徵便能達到高的辨識成功率,最後是我們使用類神經及iFuzzy LDA來判斷出腫瘤。在結果部分我們利用三個資料庫做驗證,同時請合作醫院的放射師及醫師做評估驗證,如此可以得到更正確的系統辨識成功率。論文的成功實踐將可以改善現有偵測方法的準確度,產生一個真正可以使用在輔助醫療診斷系統的乳癌偵測系統,並能有效的減少醫師判斷的時間。
According to a research report by the Department of Health, Executive Yuan, R.O.C.(Taiwan), breast cancer is the most common type of cancer in women, while the mortality rate of breast cancer of females over 40 years old is extremely high. If detected early, it can be treated early, and the mortality rate of breast cancer can be reduced.Because the gray value of the breast tumor images are higher than the other normal tissues and the texture of the breast tumor images are very obvious, we aimed at the two characteristics and propsed a algorithms for it. First, we deal the Mammography image of breast with image projection to eliminate the other images which we do not need. Next, we partition the image to many blocks which the sizes are the same. The size of each block conform the size which we need when doing feature extraction, and aim at the block to detect. If all of the pixel values in the block are zero, we will not do the characteristics extraction of it. It is express that there is no breast information in this block, and the remaining blocks characteristics extraction block. In the part of the feature extraction, we use the four characteristics to detect the breast tumor. The features are Laws’ Mask which is the representation of the texture and Histogram which is the representation of the momentum analysis of grayscale value intensity charts and GLCM and the Average Distance which is the representation of the gray value. We aimed at each ROI image block to extract the four features, but we do not use each features of the four features. According to the previous experiment, we will have high successful rate by using the feature of Laws’ Mask and the feature of modification average distance. we use neuron network and iterative fuzzy linear discriminant analysis to judge the tumor. In the part of the experiment results, we use three data base to prove our algorithm, and we also have the estimation and proof from radiographers and doctors who we cooperated. Thus, we will gain the best successful rate of the system identification. The success practicing of the project will improve the accuracy of the existed detection methods. It will produce a breast detection system to assist medical diagnosis and decrease the time of the judgment effective by doctors.
Type
thesis
File(s)
No Thumbnail Available
Name

100CSMU5585001-001.zip

Size

5.12 MB

Format

Unknown

Checksum

(MD5):1d298db50f45065d180316a4d5b1f88e

Make it your own

DSpace-CRIS can be extensively configured to meet your needs. Decide which information need to be collected and available with fine-grained security. Start updating the theme to match your Institution's web identity.

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science

  • Cookie settings
  • Privacy policy
  • End User Agreement
  • Send Feedback