Exploiting geometric information in camera networks
Abstract/Contents
- Abstract
- Ubiquitous cameras have been widely deployed on smart phones and in smart environments. Because of the richness of the visual data that they provide, the key challenges are organizing and extracting useful information from the cameras. This dissertation explores different applications utilizing geometric information in camera networks to understand the visual data. These applications include image retrieval, image organization, tracking and detection. We first present a key point selection algorithm based on geometric consistency across images. Using selected key points can significantly reduce the length of the inverted index without hurting retrieval performance. Then we show a new compact image descriptor for large scale image search. Our proposed descriptor is an extension of Vector of Locally Aggregated Descriptors (VLAD) that incorporates weak geometry information into the current framework. We demonstrate the efficiency of our method where we achieve the state-of-the-art performance on existing benchmarks. We then propose an image organization algorithm based on constraints across two domains. By solving the co-clustering problem using the alternative optimization, the algorithm exhibits the state-of-the art performance on two datasets. After that we present a real time non-rigid face tracking system deployed in smart offices based on geometric constraints on facial landmarks, which analyzes behaviors of the computer user and helps avoid potential health issues. At last, we present an occupancy detection system deployed in a smart environment which utilizes a set of filters and the geometric consistency from multiple cameras to improve the accuracy of detections.
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 | Wang, Zixuan |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering. |
| Primary advisor | Aghajan, Hamid K |
| Primary advisor | Poon, Ada Shuk Yan |
| Thesis advisor | Aghajan, Hamid K |
| Thesis advisor | Poon, Ada Shuk Yan |
| Thesis advisor | Özgür, Ayfer |
| Advisor | Özgür, Ayfer |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Zixuan Wang. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2014. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2014 by Zixuan Wang
- License
- This work is licensed under a Creative Commons Attribution 3.0 Unported license (CC BY).
Also listed in
Loading usage metrics...