A regular "boring" image of the scene. Notice the lack of surface detail on
Photo of the same scene with line features detail amplified via a projector pattern.
|Screenshot from the |
Nintendo DS game Ghostwire - ghost image overlay on the top screen.
Augmented reality is a rapidly growing research area whose main goal is to provide additional information based on extracted audio/video data from the real world. Imagine having a virtual city guide on your mobile device, able to recognize buildings from the photo and give you appropriate information, such as the history of the building. Or having a game you can play with your friends where you stream camera images of the game field to the device screen, overlay with your game avatar. Or imagine having a camera+projector system able to detect surface geometry and then project the lighting pattern which amplifies that geometry. Those are the kinds of projects we will be dealing with in this course.
You will be provided with a framework for working with our cameras and projectors (including the high-speed devices), or with a N900 smartphone platform from Nokia. Other platforms are not excluded (such as other smartphones, for example), but will not be officially supported. Most of the work will involve C++, and in case of high-speed programming knowledge of CUDA or OpenCL will be useful.
Basic knowledge in Computer Graphics, Computer Vision or Computational Photography is mandatory.
- Maximum Number of Students: 12
- Passing Criteria: Verification and Demonstration of Programs and Setup
- Registration: E-Mail to Boris Ajdin
- Students are allowed to work in groups of up to 2 people.
Next meeting: Wednesday, April 20th, 16h, room O27/3303
Potential project ideas could fall into one of the following categories:
- Projectors + cameras
- Augmented reality games
- Open Source Camera (Nokia N900)
- Android/iOS/Windows Phone 7/... apps