Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List
2259 viewed
یادگیری نیمه نظارتی و کاربرد آن در طبقه بندی تصاویر
فرج تبار، مهرداد Farajtabar, Mehrdad
Semi-supervised Learning and its Application to Image Categorization
Farajtabar, Mehrdad | 2012
1763
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 43082 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Rabiee, Hamid Reza
- Abstract:
- Traditional methods for data classification only make use of the labeled data. However, in most of the applications, labeling the unlabeled data is expensive, time consuming and requires expert knowledge. To overcome these problems, Semi-supervised Learning (SSL) methods have become an area of recent research that aim to effectively addressing the problem of limited labeled data.One of the recently introduced SSL methods is the classification based on geometric structure of the data, namely the data manifold. In this approach unlabeled data is utilized to recover the underlying structure of the data. The common assumption is that despite of being represented in a high dimensional space, data lie near a low dimensional manifold and the labeling function varies smoothly with respect to the underlying manifold. The main problem of most of the SSL methods that are based on manifold assumption is the fact that their basic implementations do not scale well to the size of data. Time and memory limitations are the major problems faced in large-scale problems. In this thesis, we investigate SSL methods with manifold assumption and try to ameliorate time and memory problems. Firstly, we propose a coarse graining algorithm to reduce the number of data while preserving the manifold structure. Then, an iterative approach to implementm anifold-based classifica- tion methods is proposed along with a theoretical analysis on the convergence issues.Finally, we evaluate the effectiveness of the proposed methods on the image categorization.The rapidly increasing number of images in the web and other databases makes image categorization an important problem. The goal of this problem is to classify the images into predefined sets in order to be used in image retrieval and analysis. Regarding the problems facing the labeling of huge number of images in the web and other databases, enhancing image cate-gorization in a semi-supervised setting seems advantageous
- Keywords:
- Semi-Supervised Learning ; Manifold ; Images Classification
محتواي پايان نامه