English  |  正體中文  |  简体中文  |  Items with full text/Total items : 66984/66984 (100%)
Visitors : 22939381      Online Users : 372
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version


    Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/77739


    Title: 運用深度學習、支持向量機及教導學習型最佳化分類糖尿病視網膜病變症狀;Diabetic Retinopathy Symptom Classification Using Deep Learning, SVM, and TLBO
    Authors: 葉乃寧;Yeh, Nai-Ning
    Contributors: 資訊工程學系
    Keywords: 糖尿病視網膜病變;深度學習;支持向量機;教導學習型最佳化;Diabetic retinopathy;Deep learning;Support vector machine;TLBO
    Date: 2018-08-16
    Issue Date: 2018-08-31 14:54:33 (UTC+8)
    Publisher: 國立中央大學
    Abstract: 糖尿病視網膜病變是長期糖尿病的併發症之一,由於糖尿病視網膜病變初期症狀很少且不易察覺,因此在早期階段很難發現。現今檢測糖尿病視網膜病變的方法,需要透過醫生用特殊光學儀器,拍攝病患眼底影像,再由醫生判斷眼底有無糖尿病視網膜病變病癥。由於醫學用特殊光學儀器非常昂貴,加上醫生判斷耗時及糖尿病視網膜病變潛在病患數量多,如果有一個自動化診斷系統,能正確診斷眼底影像是否具有視網膜病變,對一般人或病患而言,搭配手持式眼底鏡,能夠達成居家照護、自我初步診斷等;對於醫生而言,可以減少診斷糖尿病視網膜病變的時間。
    本論文中運用兩個獨立的深度卷積網路,產生每個病變期數的機率,搭配支持向量機得到最後的分類結果,與一般深度卷積網路不同的是,在池化層不是使用常見的2×2最大池化,而是使用局部最大池化,可以非整數倍且有效保留特徵,同時可以讓深度卷積網路的層數增加。實驗使用的資料庫,來自Kaggle其中一項競賽所提供的公開眼底資料庫,總共有將進九萬張影像,其中35124張影像當作訓練資料集,剩下的53572張影像當作測試資料集,在資料標籤部分,糖尿病視網膜病變可以分為兩種:非增殖性糖尿病視網膜病變跟增殖性糖尿病視網膜病變,其中非增殖性糖尿病視網膜病變依嚴重程度又可以分為三期,因此包含沒有罹患糖尿病視網膜病變的情況,標籤共有五類:正常眼底、初期非增殖性糖尿病視網膜病變、中期非增殖性糖尿病視網膜病變、後期非增殖性糖尿病視網膜病變及增殖性糖尿病視網膜病變。
    實驗結果,在分五類的情形下,本論文所提出的方法可以達到86.17%的準確率,在只分有沒有病變的情形下,可以提升至91.05%,除了演算法設計以外,還開發了一個應用程式「Deep Retina」,透過這個應用程式,使用者可以將自己的眼底影像傳送至伺服器,伺服器會將影像輸入至訓練好的深度卷積網路,再將最終結果回傳至應用程式。;Diabetic retinopathy (DR) is one of complications of long-standing diabetes, which is difficult to detect in its early stage since it only shows a few symptoms. Nowadays, the diagnosis of DR usually requires taking digital fundus images by doctors, as well as images using optical coherence tomography (OCT). Since OCT equipment is very expensive, it will benefit both the patients and the ophthalmologists if an accurate diagnosis can be made, based solely on reading digital fundus images. In the thesis, we present a novel algorithm based on deep convolutional neural network (DCNN). Unlike the traditional DCNN approach, we replace the commonly used max-pooling layers with fractional max-pooling. Two of these DCNNs with a different number of layers are trained to derive more discriminative features for classification. After combining features from meta-data of the image and DCNNs, we train a support vector machine (SVM) classifier to learn the underlying boundary of distributions of each category. For the experiments, we used the publicly available DR detection database provided by Kaggle. We used 35,124 training images and tested with 53,572 testing images. The proposed DR classifier classifies the stages of DR into five categories, labeled with an integer ranging between zero and four. The experimental results show that the proposed method can achieve a recognition rate up to 86.17%. In addition to designing a machine learning algorithm, we also develop an app called ‘Deep Retina’. Equipped with a handheld ophthalmoscope, a layman can take fundus images and perform the diagnosis automatically without intervention from ophthalmologists. It is beneficial for home care, remote medical care, and self-examination.
    Appears in Collections:[資訊工程研究所] 博碩士論文

    Files in This Item:

    File Description SizeFormat
    index.html0KbHTML27View/Open


    All items in NCUIR are protected by copyright, with all rights reserved.

    社群 sharing

    ::: Copyright National Central University. | 國立中央大學圖書館版權所有 | 收藏本站 | 設為首頁 | 最佳瀏覽畫面: 1024*768 | 建站日期:8-24-2009 :::
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback  - 隱私權政策聲明