Mobile visual search with text and image features
Abstract/Contents
- Abstract
- Visual text information is a descriptive part of many images that has been largely neglected when performing mobile visual search (MVS). In previous methods, visual text information is treated the same as any other parts of the image and is typically represented inefficiently. In this dissertation, a new way of using visual text information for MVS is presented. To use visual text information more directly, a word patch descriptor based on image gradients in a text box is developed. The word patch descriptor can be used for large scale word patch matching, and achieves a word patch matching performance that is better than the state-of-the-art image feature-based approaches. The newly developed word patch descriptor is called the word histogram of oriented gradients (Word-HOG). An image retrieval system that uses the Word-HOG descriptor for retrieving images from databases is also developed. Text retrieval tf-idf inspired scoring methods are developed for the image retrieval system. Furthermore, a random sampling-based method is used to reduce the database. The image retrieval system achieves comparable performance to state-of-the-art image feature-based retrieval systems for images of book covers, and performs better than state-of-the-art text-based retrieval systems for images of book pages. Lastly, a lossy compression method is developed for the Word-HOG descriptor and is used with the Word-HOG-based image retrieval system to construct an MVS system. The system achieves more than 10-to-1 query size reduction for images of book covers while achieving more than 16-to-1 query size reduction for images of book pages.
Description
Type of resource | text |
---|---|
Form | electronic; electronic resource; remote |
Extent | 1 online resource. |
Publication date | 2014 |
Issuance | monographic |
Language | English |
Creators/Contributors
Associated with | Tsai, Shang-Hsuan |
---|---|
Associated with | Stanford University, Department of Electrical Engineering. |
Primary advisor | Girod, Bernd |
Thesis advisor | Girod, Bernd |
Thesis advisor | Gray, Robert M, 1943- |
Thesis advisor | Grzeszczuk, Radek, 1967- |
Advisor | Gray, Robert M, 1943- |
Advisor | Grzeszczuk, Radek, 1967- |
Subjects
Genre | Theses |
---|
Bibliographic information
Statement of responsibility | Sam S. Tsai. |
---|---|
Note | Submitted to the Department of Electrical Engineering. |
Thesis | Thesis (Ph.D.)--Stanford University, 2014. |
Location | electronic resource |
Access conditions
- Copyright
- © 2014 by Shang-Hsuan Tsai
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...