Camera-based human computer interaction

• Main Author: Ong, Da Wei.
• Other Authors: Chan Kap Luk
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Home entertainment systems
• Online Access: http://hdl.handle.net/10356/17976
• Institution: Nanyang Technological University

Camera-based Human Computer Interaction (HCI) is a study of computer vision based strategies to detect and extract human body motion and gestures from a video camera. These detected movements are then translated into computer commands. Camera-based HCI can be implemented in 3D computer games, robot control and many other applications. This report shall first discuss ways to detect and track the positions of the hand or the marker that represents the human hand. In this part of the report, various detection methods and tracking methods will be discussed. 3D object detection by stereo vision and multiple object detection methods will also be discussed in this report. The second objective of this report is to discuss various interfaces to interact the user with the computer through the use of camera-based HCI. The aim is to make the interfaces interactive and fun. As the developed program is written in C++ and running in a Microsoft Windows XP environment, this report shall also investigate various ways to send a computer input to Windows from the C++ program. Interfaces such as the mouse simulation and keyboard simulation interfaces will be discussed. These interfaces are tested on various popular game demos such as Core of Duty Demo and Need for Speed Prostreet Demo. Other advanced interfaces such as the pointer interface and the multi-touch interface are also discussed.