Interframe compression of visual feature descriptors for mobile augmented reality
Abstract/Contents
- Abstract
- Streaming mobile augmented reality applications require both real-time recognition and tracking of objects of interest in a video scene and many of these applications utilize local image features. Typically, local feature descriptors are calculated from the gradients of a canonical patch around a keypoint in the image. For mobile applications where data are transmitted over a network for feature detection and image matching, it is desirable that the amount of data sent is as low as possible to reduce the latency and the cost of the system. We first study the possible architectures for a system that uses a single image, and propose a novel architecture based on encoding canonical patches extracted on the mobile phone. Our system represents an intermediate solution between a system that sends the whole image to the server, and a system that extracts the feature descriptors on the phone. We then consider the problem of low bit-rate streaming mobile augmented reality, and present a two-stage solution. We propose a temporally coherent keypoint detector and illustrate its efficiency through the evaluation of the similarity between the temporally coherent canonical patches. Moreover, we design efficient interframe predictive coding techniques for canonical patches, feature descriptors and keypoint locations. In the proposed system, we strive to transmit each patch or its equivalent feature descriptor with as few bits as possible by simply modifying a previously transmitted patch or descriptor. Our solution enables server-based mobile augmented reality where a continuous stream of salient information, sufficient for image-based retrieval and object localization, can be sent over a wireless link at a bit-rate of about 20 - 30 kbps, which is practical for today's wireless links and less than one-tenth of the bit-rate needed to stream the whole video to the server.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2013 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Makar, Mina |
|---|---|
| Associated with | Stanford University, Department of Electrical Engineering. |
| Primary advisor | Girod, Bernd |
| Thesis advisor | Girod, Bernd |
| Thesis advisor | Apostolopoulos, John G |
| Thesis advisor | El Gamal, Abbas A |
| Advisor | Apostolopoulos, John G |
| Advisor | El Gamal, Abbas A |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Mina Makar. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2013. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2013 by Mina Ayman Saleh Yanni Makar
- 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...