Loading...

Disease Classification Based on Graph Learning using fMRI Datasets

Arasteh, Ali | 2023

90 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56237 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Amini, Arash
  7. Abstract:
  8. In the past few years, the available knowledge in graph-based processing has made significant progress, and as a result, powerful tools have been created. In this regard, graph learning with the assumption of data smoothness on the final result can be considered a successful example. Briefly, in graph learning, to describe the relationship between the problem components, a graph is learned using the available data whose nodes represent the problem components, and its edges represent how much these components are connected. The usefulness of this method lies in the possibility of using the obtained graph as the input to currently known methods of classification and achieving better results due to having a deeper understanding of the data. The main goal of this project is to re-examine two medical datasets, ABIDE and Development, considering this relatively new approach. In the ABIDE dataset, a collection of fMRI data related to people with ASD and healthy people has been collected. In the past, many attempts have been made to classify ASD on this dataset. In most cases, the proposed approach is to form a functional connectivity graph from fMRI data using traditional graph construction methods (usually calculation of Pearson's correlation) and apply different classifiers on obtained graphs. In this research, firstly, the performance of the proposed methods of smooth graph learning is compared to the previous approaches. Then, by changing the optimization problem related to graph learning, another approach is presented to form the functional connectivity graph, and its performance is evaluated. Also, on the Development dataset, the performance of traditional graph construction methods and methods based on graph learning by solving an optimization problem is compared to separate children and adults. In the mentioned cases, the evaluation and comparison criterion is the accuracy percentages obtained for the classification of ASD or the separation of children and adults by applying similar methods on the graphs obtained from different approaches. The result of this research is the superiority of new graph learning methods over previous approaches in both datasets. This superiority is about 8% in the ABIDE dataset and about 20% in the Development dataset over the average accuracy percentages
  9. Keywords:
  10. Graph Learning ; Atrial Aeptal Defect (ASD) ; Functional Magnetic Resonance Imaging (FMRI) ; Disease Classification ; Graph-Based Learning

 Digital Object List

 Bookmark

  • مقدمه
    • تعریف مسئله
    • اهمیت موضوع
    • اهداف پژوهش
    • ساختار پایان‌نامه
  • مفاهیم اولیه
    • گراف
      • ‌تعریف‌های مختلف گراف
      • کاربردهای گراف در حوزه‌های مختلف
    • تصویربرداری تشدید مغناطیسی کارکردی (fMRI)
    • اطلس‌های مغزی
    • ارتباطات مغزی
    • طبقه‌بندی و معیارهای ارزیابی آن
      • تعریف مسئله طبقه‌بندی
      • معیارهای ارزیابی کیفیت طبقه‌بندی
  • مروری بر ادبیات
    • یادگیری گراف با فرض هموار بودن داده‌های موجود روی آن
      • مسئله‌های بهینه‌سازی پیشنهاد شده
      • دو مشکل مسئله‌های بهینه‌سازی پیشنهاد شده و راه‌حل مقابله با هر یک
    • دیتاست ABIDE
      • معرفی پروژه ABIDE
      • رویکردهای پیشین برای طبقه‌بندی ASD روی دیتاست ABIDE
    • دیتاست Development
  • نتایج
    • روش‌های پیشنهادی
    • دیتاست ABIDE
      • گراف میانگین به عنوان طبقه‌بندی ساده
      • تغییر تعریف ماتریس فاصله در مسئله بهینه‌سازی
      • بررسی اطلس‌های مغزی با منطقه‌های مورد علاقه کمتر و بیشتر
      • بررسی عملکرد طبقه‌بند SVM
      • بررسی رویکرد Leave-One-Out
      • بررسی نمونه‌های مکان‌های مختلف به طور جداگانه
    • دیتاست Development
      • مقایسه روش‌های پیشین ساخت گراف با روش‌های جدید یادگیری گراف
      • دو اطلس مغزی دیگر
      • مصورسازی برتری روش‌های جدید یادگیری گراف
      • مقایسه گراف میانگین گروه کودکان با گراف میانگین گروه بزرگسالان
  • جمع‌بندی
  • مراجع
  • مطالب تکمیلی
    • اثبات قضیه 3-1
    • اثبات قضیه 3-2
    • اثبات قضیه 3-3
...see more