跳到主要內容

臺灣博碩士論文加值系統

(3.236.110.106) 您好!臺灣時間:2021/07/26 00:13
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:黃圓媚
研究生(外文):Yuan-May Huang
論文名稱:建構多媒體電子病歷與其影像特徵比對系統之可能利用
論文名稱(外文):Building Medical Record and Investigating the Usage of Image Matching Technique with Medical Record
指導教授:劉德明劉德明引用關係
指導教授(外文):Der-Ming Liou
學位類別:碩士
校院名稱:國立陽明大學
系所名稱:衛生資訊與決策研究所
學門:醫藥衛生學門
學類:公共衛生學類
論文種類:學術論文
論文出版年:2004
畢業學年度:92
語文別:英文
論文頁數:56
中文關鍵詞:主成份分析小波轉換大腸直腸鏡特徵辦識
外文關鍵詞:PCAWavelet TransformSVMColonoscopyPattern Recognition
相關次數:
  • 被引用被引用:2
  • 點閱點閱:127
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
由於醫院內每日門診量數目龐大,又以外科門診更需要從初診進而轉入內診,並運用先進醫療儀器的診察,耗費的時間和人力因此也會跟著增加。針對直腸外科大腸鏡室而言,使用大腸鏡診察病況的病患所產生的影像數量,必須要有效的管理和儲存,在診斷病情以及參詳對照將會有事半功倍的效果。如果能夠增加快速瀏覽影像的功能,並且將可疑部分標出,作為醫師在診斷病情上的參考,以增進效率;因此,運用一些方法,來探討何種方式的判斷能夠增加機器對有病徵影像更好的辨識率。
本研究收集直腸外科內為病患拍攝疑似為腫瘤的各類影像;其中包括CT,切片和大腸內視鏡等影像,利用Web的介面,很容易地將所有相關影像管理存放,並且使用了直腸外科常用字彙建立metadata,來連結各個相關的影像。並且針對大腸內視鏡在拍攝病患腸道中,疑似為腫瘤的影像,利用影像處理的技術以及資料探勘中的分類方式,試著探討是否能夠將腫瘤以及非腫瘤有效的分類,既而達到協助醫師能夠很快速的了解該張影像的重點所在位置。
本篇論文利用support vector machine(SVM)進行資料的分類,並且算出準確率、敏感度及特異度,作為分類好壞的依據;實驗結果使用小波轉換之後再加上主成份分析的準確率為96.296 %,敏感度及特異度則分別為0.975、0.987。
Due to the number of increasing outpatients more and more every day in hospital, especially surgery department, it had been a long time from first consultation to using endoscope, and nurses and doctors would keep their mind on diagnosing. As far as Rectal and Colon Surgery Department concerned, images were taken in every diagnosis, so we must manage these medical images well. The effect was more efficient after managing database. We could try to process medical images before doctors’ viewing since this method could decrease the viewing time, mark the doubtable position, and be a reference for diagnosis. Therefore by using some image processing, we would be able to investigate what process could be more efficient in machine recognition.
In this research, we collected the images that be doubted of tumor existence for patients, including CT, specimen and colonoscopy photos. We designed a user interface on the Web that could add new records, edit and delete records and search the images, and construct metadata for doctors’ contribution to connect the images to search easily. Besides, we focused on doubting images which can be taken in patients’ colon and use image processing and classification in Data Mining. We can use that to explore weather classify tumor and non-tumor. We expect the result can be hopeful to categorization.
In this thesis, the results of classification using support vector machine (SVM) were suitable. Gold standard experts based on experienced physician. Accuracy is 96.296 %, sensitivity is 0.975 and specificity is 0.987 using wavelet transform + Principal Component Analysis (PCA).
[1] American Cancer Society. Cancer Facts and Figure. 2004.
[2] Depaertment of Health EYTROC. Cause of Death Statistis. 2004.
[3] Rex DK, Weddle RA, Lehman GA et al. Flexible sigmoidoscopy plus air contrast barium enema versus colonoscopy for suspected lower gastrointestinal bleeding. Gastroenterology 1990 Apr;98(4):855-61.
[4] Jain AK, Duin RPW, Jianchang M. Statistical pattern recognition: a review. Pattern Analysis and Machine Intelligence, IEEE Transactions on 2000;22(1):4-37.
[5] Taur JS. Medical image compression using principal component analysis. 1996 p. 903-6.
[6] Kusumoputro B. Neural network diagnosis of malignant skin cancers using principal component analysis as a preprocessor. 1998 p. 310-5.
[7] D.L.Donho. wedgelets: Nearly-minimax estmation of edges. 1999.
[8] Todorovic S, Nechyba MC. Multiresolution linear discriminant analysis: efficient extraction of geometrical structures in images. 2003 p. I-32.
[9] JIA-WEI PAN. The application of The Depth-from-Focus Techniques 2004.
[10] Sujan VA, Mulqueen MP. Image identification using the segmented Fourier transform and competitive training in the HAVNET neural network. 2001 p. 489-92.
[11] Karkanis SA. Computer-aided tumor detection in endoscopic video using color wavelet features. Information Technology in Biomedicine, IEEE Transactions on 2003;7(3):141-52.
[12] Buciu I, kotropoulos C, Pitas I. ICA and Gabor representation for facial expression recognition. 2003 p. II-8.
[13] Vapnik. The Nature of Statistical of Learning Theory. 1995.
[14] Vapnik, S.Golowich, A.Smola. Support vector method for function approximation, regression estimation, and signal processing. 1997.
[15] Mousa WAH, Khan MAU. Lung nodule classification utilizing support vector machines. 2002 p. III-153.
[16] Romero E, Raymackers JM, Macq B, Cuisenaire O. Automatic fibrosis quantification by using a k-NN classificator. 2001 p. 2609-12.
[17] Oh-Wook Kwon, Kwokleung Chan, Jiucang Hao, Te-Won Lee. Emotion Recognition by Speech Signals. 2002.
[18] C.-W.Hsu, C.-J.Lin. A comparison of methods for multi-class support vector machines. IEEE Transactionson Neural Networks 2002.
[19] C.-J.Lin. SVM reasearch. 2004.
[20] P.-H.Chen, C.-J.Lin, B.Scholkopf. A tutorial on nu-support vector machines 2003.
[21] C.-C.Chang, C.-W.Hsu, C.-J.Lin. A practical guide to support vector classification 2003.
[22] Van de Wouwer G, Scheunders P, Van Dyck D. Statistical texture characterization from discrete wavelet representations. Image Processing, IEEE Transactions on 1999;8(4):592-8.
[23] Jean-Philippe Vert. Introduction to support vector machines applications to computational biology 2001.
[24] M.-W.Chang, C.-J.Lin. Leave-one-out Bounds for Support Vector Regression Model Selection 2003.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top