CS 6751

Winter 1999

Homework

• Novel input and output device

Vision-based input device

 
Computer vision has been around for quite a while. Early research work has been applied in artificial intelligence and robotics. However, recently much work has been explored to use vision as a new way to interpret and understand environment around us. Thanks to faster processing power and better quality of digital images, this idea will perhaps be fully realized in the near future. Research in computer vision has helped us solving many difficult problems. Examples include navigation and planning in robotics, modeling from images in computer graphics, registration and calibration in augmented reality and many more. An interesting problem now is how we can apply vision to make computers effectively perceive, interpret and understand us.

 I think that a main issue in designing vision-based interface for everyday environment is to make it visible and easy to learn to operate. I do not mean that everyone who wants to use the system will have to understand how the implementation works. However, designers need to keep in mind that a user must be able to create a mental model that will correctly map to a system image. I do not mean that every one needs to know its implementation. The system must provide enough feedback to users so that they can tell how reliable the interpretation is or whatever result provided by a vision system. In other words, users are the ones who decide whether to trust the system or not.

 Links to publication and demo from various places

• Augmenting human sensory, integration of vision and sensing

• Wearable computing page at Georgia Tech

• Vision and Modeling Group, one of their main goals is to make computers perceive and understand everyday environment around us

• Affective Computing, trying to bridge the gap between computational system and human emotions

• Many projects at Computational Perception Lab here at Georgia Tech, real-time analysis and tracking of lab inhabitants in a real office environment

By combining tracking technology with computer graphic, we have a new tool to interact with computer generated environment. For example, in a 3D navigation system or walkthrough of the environment, real time tracking can give us position and pose estimation of a user. We then use the data in creating accurate response to the user’s perception. Some may say that this is a form of virtual reality, but I would like to think of it more general than that. I think that it should include augmented reality as well. In augmented reality, a major goal is to combine computer generated object with the real environment. Thus, the user may want to operate in the real environment with added objects. In this case, those added objects must seamlessly blend in. Either way, a goal of this system should try to ensure that the user can fully immerse into the environment and correctly perceive it.

• A survey of immersive technology

• USC Integrated Medias System Center ,research on immersive environment, including augmented reality